Intelligent generation method and system for industrial energy-saving renovation solutions

Abstract

This invention provides an intelligent generation method and system for industrial energy-saving retrofit schemes, relating to the field of industrial energy-saving technology. The method includes: acquiring operating parameters and energy consumption parameters of industrial system equipment and constructing equipment topology relationships and dynamic load characteristics; identifying key energy consumption paths based on coupling dependency and analyzing equipment efficiency evaluation values ​​to obtain retrofit potential values; tracing back to the source equipment along the energy consumption impact transmission chain and obtaining retrofit adaptability based on fluctuation characteristics and adjustment capabilities to screen target retrofit equipment; then constructing an operating state constraint space optimization to obtain energy-saving operating parameters and propagating them forward along the transmission chain to generate collaborative retrofit configurations, thus realizing intelligent and precise generation and efficient collaborative optimization of energy-saving retrofit schemes.

Description

Technical Field

[0001] This invention relates to the field of industrial energy conservation technology, and in particular to an intelligent method and system for generating industrial energy conservation retrofit schemes. Background Technology

[0002] In the field of industrial energy-saving retrofitting, retrofitting plans are typically formulated based on energy consumption statistics or offline test efficiency of individual equipment. By collecting parameters such as equipment operating power and start-up / shutdown time, combined with the equipment's nameplate rated efficiency or empirical thresholds, it is directly determined whether the equipment belongs to a high-energy-consuming or low-efficiency segment, and replacement or upgrade recommendations are made accordingly. Some plans rank equipment according to its historical total energy consumption as a percentage, prioritizing the equipment with the highest energy consumption, or use simple linear regression models to predict energy-saving potential. These plans rely heavily on static parameters and isolated analyses, lacking a systematic consideration of the operational coupling relationships between equipment.

[0003] The main drawback of existing technologies lies in neglecting the energy transfer and dynamic load correlation between equipment. In industrial systems, equipment is interconnected through processes or energy media. Efficiency fluctuations in one piece of equipment often affect the actual operating status of upstream and downstream equipment. Simply assessing the efficiency of a single machine cannot identify the overall energy consumption deviation caused by this coupling dependency. For example, a decrease in the efficiency of a compressor may force downstream equipment to increase its power to maintain output, but traditional technologies cannot trace the impact path across equipment, leading to modification suggestions focusing on localized aspects and missing opportunities for system-level optimization.

[0004] Furthermore, existing technologies lack time-domain decomposition of load dynamic characteristics. In actual industrial processes, equipment loads often exhibit high-frequency fluctuations and periodic changes. Existing technologies tend to use average or peak values ​​as the basis for modification, failing to analyze the load distribution characteristics over time. This makes it difficult for the evaluation results to reflect the true efficiency of the equipment under different load segments, limiting the universality and adaptability of modification schemes. In particular, when dealing with sudden load changes, it may lead to ineffective or excessive modifications. Summary of the Invention

[0005] This invention provides an intelligent generation method and system for industrial energy-saving renovation schemes, which can at least solve some of the problems existing in the prior art.

[0006] A first aspect of this invention provides an intelligent generation method for industrial energy-saving retrofit schemes, comprising:

[0007] The operating parameters and energy consumption parameters of each energy-consuming device in the industrial system are obtained. The operating parameters are decomposed in the time domain to obtain dynamic load characteristics. Based on the energy consumption parameters, the device topology relationship is constructed and correlated with the dynamic load characteristics to obtain the device efficiency evaluation value and coupling dependency.

[0008] Based on the coupling dependency, key energy consumption paths are identified and propagation analysis is performed on the equipment efficiency evaluation values ​​to obtain the energy consumption impact transmission chain. Based on the equipment efficiency evaluation values ​​of each node in the energy consumption impact transmission chain, adjustability is determined to obtain the transformation potential value.

[0009] Based on the transformation potential value, the energy consumption source node is identified by reverse tracing along the energy consumption impact transmission chain. The fluctuation characteristics are determined based on the dynamic load characteristics corresponding to the energy consumption source node and matched with the pre-configured equipment adjustment capabilities to obtain the transformation adaptability. The target transformation equipment is selected based on the transformation adaptability.

[0010] Based on the dynamic load characteristics and preset process constraints, the operating state constraint space corresponding to the target modification equipment is constructed, and the energy consumption parameters are optimized and solved to obtain energy-saving operating parameters. The energy-saving operating parameters are forward propagated along the energy consumption influence transmission chain to calculate the parameter adjustment amount of each associated equipment. Based on the parameter adjustment amount and the equipment topology relationship, a collaborative modification configuration is generated and a modification scheme is output.

[0011] In one alternative implementation,

[0012] The system acquires the operating parameters and energy consumption parameters of each energy-consuming device in the industrial system. It then performs time-domain decomposition on the operating parameters to obtain dynamic load characteristics. Based on the energy consumption parameters, it constructs device topology relationships and performs correlation calculations with the dynamic load characteristics to obtain device efficiency evaluation values ​​and coupling dependencies, including:

[0013] The operating parameters and energy consumption parameters of each energy-consuming device in the industrial system are obtained. The operating parameters are segmented according to a preset time window and frequency domain transformation is performed to obtain spectral components. The dominant frequency components and amplitude distribution are extracted from the spectral components and combined to obtain dynamic load characteristics.

[0014] Based on the energy input-output relationship in the energy consumption parameters, identify the energy transfer direction between each energy-consuming device, construct a device topology graph by taking each energy-consuming device as a node and the energy transfer direction as a directed edge, and determine the device topology relationship based on the device topology graph;

[0015] The dynamic load characteristics are arranged into a load characteristic matrix according to the energy-consuming equipment. A topological adjacency matrix is ​​constructed based on the equipment topology relationship. The load characteristic matrix and the topological adjacency matrix are multiplied to obtain the neighborhood propagation characteristics. The deviation between the neighborhood propagation characteristics and the load characteristic matrix is ​​calculated and normalized to obtain the equipment efficiency evaluation value.

[0016] Extract the upstream and downstream device sets of each energy-consuming device from the device topology and calculate the number of energy-dependent paths. Then, weight and sum the number of energy-dependent paths with the device efficiency evaluation value to obtain the device importance. Based on the device importance, identify key dependent devices and calculate the coupling dependency based on the connection strength of the key dependent devices in the device topology.

[0017] In one alternative implementation,

[0018] Based on the coupling dependency, key energy consumption paths are identified, and propagation analysis is performed on the equipment efficiency evaluation values ​​to obtain the energy consumption impact transmission chain. Adjustability is determined based on the equipment efficiency evaluation values ​​of each node in the energy consumption impact transmission chain to obtain the modification potential value, including:

[0019] The connection paths between each energy-consuming device are extracted from the device topology. The cumulative coupling dependency value of each node on each connection path is calculated. Based on the cumulative coupling dependency value, all connection paths are sorted and filtered to obtain the key energy-consuming paths.

[0020] The equipment efficiency evaluation values ​​are arranged in the order of nodes in the key energy consumption path to form an initial efficiency vector. A path propagation matrix is ​​constructed based on the equipment topology relationship, and a first-order propagation efficiency vector is calculated by combining the initial efficiency vector with the path propagation matrix. An iterative matrix multiplication operation is performed on the first-order propagation efficiency vector and the path propagation matrix, and the results are accumulated to obtain a multi-order propagation efficiency vector. An energy consumption influence transmission chain is constructed based on the cumulative efficiency propagation value of each node in the multi-order propagation efficiency vector.

[0021] The efficiency sensitivity is obtained by extracting the equipment efficiency assessment value of each node from the energy consumption impact transmission chain and the corresponding efficiency propagation cumulative value in the multi-order propagation efficiency vector, calculating the ratio between the equipment efficiency assessment value and the efficiency propagation cumulative value, extracting the load fluctuation range of the corresponding equipment of each node from the dynamic load characteristics and calculating the adjustment response degree in combination with the efficiency sensitivity, and determining the adjustability based on the adjustment response degree and generating the transformation potential value.

[0022] In one alternative implementation,

[0023] Based on the transformation potential value, the energy consumption source nodes are identified by reverse tracing along the energy consumption impact transmission chain. The fluctuation characteristics are determined based on the dynamic load characteristics corresponding to the energy consumption source nodes and matched with the pre-configured equipment adjustment capabilities to calculate the transformation adaptability. Based on the transformation adaptability, target transformation equipment is selected and determined, including:

[0024] The potential values ​​of the transformation are arranged into potential vectors according to the nodes in the energy consumption impact transmission chain. A back propagation matrix is ​​constructed based on the equipment topology. The importance weight of the nodes corresponding to the potential vector is determined based on the back attention mechanism. The weighted back tracing vector is obtained by combining the back propagation matrix. The energy contribution gradient is extracted from the weighted back tracing vector and the node corresponding to the gradient vanishing position is identified as the energy consumption source node.

[0025] The load time series data of the equipment corresponding to the energy source node is extracted from the dynamic load characteristics. Wavelet packet decomposition is performed on the load time series data to obtain fluctuation components of different frequency bands. The dominant frequency band is identified from the fluctuation components and instantaneous energy is extracted. An energy envelope curve is constructed based on the instantaneous energy and the peak distribution pattern is extracted to obtain the fluctuation characteristics.

[0026] The system acquires the pre-configured equipment adjustment capability and extracts the adjustment response bandwidth and energy buffer capacity. It extracts the frequency domain response requirement and instantaneous power surge from the fluctuation characteristics. It performs matching calculations between the frequency domain response requirement and the adjustment response bandwidth to obtain the frequency adaptation coefficient. It performs matching calculations between the instantaneous power surge and the energy buffer capacity to obtain the capacity adaptation coefficient. It then combines the frequency adaptation coefficient to solve for the modification adaptation degree. Based on the modification adaptation degree, it screens and determines the target modification equipment.

[0027] In one alternative implementation,

[0028] The transformation potential values ​​are arranged into potential vectors according to the nodes in the energy consumption impact transmission chain. A backpropagation matrix is ​​constructed based on the equipment topology. The node importance weights corresponding to the potential vectors are determined based on the back attention mechanism, and the weighted back tracing vector is obtained by combining the backpropagation matrix.

[0029] Extract the transformation potential value of each node from the energy consumption impact transmission chain and arrange them into a potential vector according to the node order. Extract the upstream dependency relationship between each energy-consuming device from the device topology relationship. Reverse the energy transmission direction in the upstream dependency relationship and construct a reverse connection graph. Calculate the node reverse propagation coefficient based on the in-degree and out-degree of each node in the reverse connection graph and fill it into a reverse propagation matrix according to the node position.

[0030] The potential vector is used as the query vector, and the cumulative efficiency propagation value of each node is extracted from the energy consumption influence transmission chain as the key vector. The query vector and the key vector are multiplied by a dot product and a scaling transformation is applied to obtain the attention score. The attention score is normalized to obtain the attention weight distribution. The potential vector is weighted and summed based on the attention weight distribution to obtain the context-aware potential vector. The context-aware potential vector and the potential vector are residually connected and adjusted by a gating unit to obtain the node importance weight.

[0031] The node importance weights are converted into diagonal matrices and multiplied with the backpropagation matrix to obtain a weighted backpropagation matrix. The potential vector is then multiplied with the weighted backpropagation matrix using iterative matrix multiplication, and the results are accumulated to obtain a weighted backtracking vector.

[0032] In one alternative implementation,

[0033] Based on the dynamic load characteristics and preset process constraints, an operating state constraint space corresponding to the target modified equipment is constructed, and the energy consumption parameters are optimized and solved to obtain energy-saving operating parameters, including:

[0034] Based on the dynamic load characteristics, the load change rate is determined. The process flow stability boundary and capacity guarantee threshold are extracted from the preset process constraints. The dynamic constraint boundary is determined based on the load change rate and the process flow stability boundary, and the static constraint boundary is determined based on the capacity guarantee threshold. The intersection operation of the dynamic constraint boundary and the static constraint boundary is performed to obtain the feasible operating domain.

[0035] Based on the feasible operating domain, a set of candidate operating points is divided and the energy input to output efficiency ratio of each candidate operating point is calculated to obtain an initial efficiency score. Downstream related devices are extracted from the energy consumption influence transmission chain and the energy consumption transmission influence of each candidate operating point on the downstream related devices is calculated. The energy consumption transmission influence and the initial efficiency score are weighted and summed to obtain a global efficiency score.

[0036] Based on the global efficiency score, a subset of high-efficiency operating points is obtained. The operating points in the subset of high-efficiency operating points are time-series aligned with the dynamic load characteristics, and the time offset is calculated to obtain the response delay coefficient. The global efficiency score and the response delay coefficient are normalized and weighted to obtain a comprehensive evaluation index. Based on the comprehensive evaluation index, the optimal operating point is selected and the corresponding control parameter combination is extracted to obtain energy-saving operating parameters.

[0037] In one alternative implementation,

[0038] The energy-saving operating parameters are forward-propagated along the energy consumption impact transmission chain to calculate the parameter adjustment amount for each associated device. Based on the parameter adjustment amount and the device topology relationship, a collaborative transformation configuration is generated and a transformation plan is output, including:

[0039] Extract the control parameter combination and energy input change value of the target modified equipment from the energy-saving operation parameters, extract the propagation direction and energy transfer coefficient in the energy consumption influence transmission chain, propagate the energy input change value to downstream related equipment layer by layer and accumulate to obtain the energy consumption change of each related equipment, and perform reverse calculation based on the energy consumption change and the equipment efficiency evaluation value of each related equipment to obtain the parameter adjustment amount of each related equipment.

[0040] The parallel and series relationships between devices are extracted from the device topology. Based on the parallel relationship, the parameter adjustment amount in the parallel device group is load balanced and distributed to obtain the balanced adjustment amount. Based on the series relationship, the parameter adjustment amount in the series device chain is time-series coordinated and corrected to obtain the coordinated adjustment amount.

[0041] The control parameters of the target equipment to be modified, along with the equalization adjustment amount and the coordination adjustment amount, are combined according to the equipment hierarchy in the equipment topology to construct a collaborative modification configuration and output a modification scheme.

[0042] A second aspect of this invention provides an intelligent generation system for industrial energy-saving retrofit schemes, comprising:

[0043] The coupling calculation unit is used to obtain the operating parameters and energy consumption parameters of each energy-consuming device in the industrial system, perform time-domain decomposition on the operating parameters to obtain dynamic load characteristics, construct the device topology relationship based on the energy consumption parameters, and perform correlation calculation with the dynamic load characteristics to obtain the device efficiency evaluation value and coupling dependency.

[0044] The propagation analysis unit is used to identify key energy consumption paths based on the coupling dependency and perform propagation analysis on the equipment efficiency evaluation value to obtain the energy consumption impact transmission chain. Based on the equipment efficiency evaluation value of each node in the energy consumption impact transmission chain, the unit performs adjustability determination to obtain the transformation potential value.

[0045] The source matching unit is used to identify energy consumption source nodes by tracing back along the energy consumption impact transmission chain based on the transformation potential value, determine the fluctuation characteristics based on the dynamic load characteristics corresponding to the energy consumption source nodes and match them with the pre-configured equipment adjustment capabilities to calculate the transformation adaptability, and screen and determine the target transformation equipment based on the transformation adaptability.

[0046] The scheme generation unit is used to construct the operating state constraint space corresponding to the target modification equipment based on the dynamic load characteristics and preset process constraints, optimize and solve the energy consumption parameters to obtain energy-saving operating parameters, forward propagate the energy-saving operating parameters along the energy consumption influence transmission chain to calculate the parameter adjustment amount of each associated equipment, generate a collaborative modification configuration based on the parameter adjustment amount and the equipment topology relationship, and output the modification scheme.

[0047] A third aspect of the present invention provides an electronic device, comprising:

[0048] A processor and a memory for storing processor-executable instructions, wherein the processor is configured to invoke instructions stored in the memory to perform the aforementioned method.

[0049] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0050] In this invention, based on the time-domain decomposition of industrial equipment operating parameters and energy consumption parameters, and the correlation calculation of equipment topology relationships, the efficiency assessment value and coupling dependency of each piece of equipment can be accurately quantified. This allows for the accurate identification of high-energy-consuming links and the mutual influence relationships between equipment within the system, providing a reliable data foundation for subsequent upgrades. Through the analysis of coupling dependencies, key energy-consuming paths can be automatically discovered, and efficiency assessment values ​​can be propagated along these paths to quickly locate nodes with the greatest energy-saving potential. Simultaneously, by combining adjustability determination, the upgrade potential value is obtained, avoiding blind upgrades of inefficient equipment and significantly improving the economy and feasibility of the upgrade plan.

[0051] By tracing the energy consumption impact chain backwards to the energy consumption source node and matching the dynamic load fluctuation characteristics of that node with the equipment's adjustment capabilities, the target equipment best suited to the system's operating state can be selected. Based on the differentiated matching of fluctuation characteristics and adjustment capabilities, it is ensured that the upgraded equipment can effectively respond to load changes and operate stably under existing process conditions, avoiding over-investment or system instability caused by improper equipment selection, and significantly reducing upgrade costs and technical risks.

[0052] Optimizing energy consumption parameters within the operational constraint space generates optimal energy-saving operating parameters that conform to process constraints. By calculating parameter adjustments for each associated device through forward propagation, the overall retrofit plan is ensured to be coordinated and consistent. The collaborative retrofit configuration generated based on device topology relationships automatically outputs a complete retrofit plan including multi-device commissioning parameters and equipment replacement schemes. This achieves global energy-saving optimization from source to end, minimizing the workload of post-retrofit debugging and improving the intelligence and implementation efficiency of industrial energy-saving retrofits. Attached Figure Description

[0053] Figure 1 This is a flowchart illustrating the intelligent generation method for industrial energy-saving renovation schemes according to an embodiment of the present invention.

[0054] Figure 2 This is a flowchart illustrating the energy consumption source tracing process of the intelligent generation method for industrial energy-saving renovation schemes according to an embodiment of the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0056] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0057] Figure 1 This is a flowchart illustrating the intelligent generation method for industrial energy-saving retrofit schemes according to an embodiment of the present invention. Figure 1 As shown, the method includes:

[0058] The operating parameters and energy consumption parameters of each energy-consuming device in the industrial system are obtained. The operating parameters are decomposed in the time domain to obtain dynamic load characteristics. Based on the energy consumption parameters, the device topology relationship is constructed and correlated with the dynamic load characteristics to obtain the device efficiency evaluation value and coupling dependency.

[0059] Based on the coupling dependency, key energy consumption paths are identified and propagation analysis is performed on the equipment efficiency evaluation values ​​to obtain the energy consumption impact transmission chain. Based on the equipment efficiency evaluation values ​​of each node in the energy consumption impact transmission chain, adjustability is determined to obtain the transformation potential value.

[0060] Based on the transformation potential value, the energy consumption source node is identified by reverse tracing along the energy consumption impact transmission chain. The fluctuation characteristics are determined based on the dynamic load characteristics corresponding to the energy consumption source node and matched with the pre-configured equipment adjustment capabilities to obtain the transformation adaptability. The target transformation equipment is selected based on the transformation adaptability.

[0061] Based on the dynamic load characteristics and preset process constraints, the operating state constraint space corresponding to the target modification equipment is constructed, and the energy consumption parameters are optimized and solved to obtain energy-saving operating parameters. The energy-saving operating parameters are forward propagated along the energy consumption influence transmission chain to calculate the parameter adjustment amount of each associated equipment. Based on the parameter adjustment amount and the equipment topology relationship, a collaborative modification configuration is generated and a modification scheme is output.

[0062] In one alternative implementation,

[0063] The system acquires the operating parameters and energy consumption parameters of each energy-consuming device in the industrial system. It then performs time-domain decomposition on the operating parameters to obtain dynamic load characteristics. Based on the energy consumption parameters, it constructs device topology relationships and performs correlation calculations with the dynamic load characteristics to obtain device efficiency evaluation values ​​and coupling dependencies, including:

[0064] The operating parameters and energy consumption parameters of each energy-consuming device in the industrial system are obtained. The operating parameters are segmented according to a preset time window and frequency domain transformation is performed to obtain spectral components. The dominant frequency components and amplitude distribution are extracted from the spectral components and combined to obtain dynamic load characteristics.

[0065] Based on the energy input-output relationship in the energy consumption parameters, identify the energy transfer direction between each energy-consuming device, construct a device topology graph by taking each energy-consuming device as a node and the energy transfer direction as a directed edge, and determine the device topology relationship based on the device topology graph;

[0066] The dynamic load characteristics are arranged into a load characteristic matrix according to the energy-consuming equipment. A topological adjacency matrix is ​​constructed based on the equipment topology relationship. The load characteristic matrix and the topological adjacency matrix are multiplied to obtain the neighborhood propagation characteristics. The deviation between the neighborhood propagation characteristics and the load characteristic matrix is ​​calculated and normalized to obtain the equipment efficiency evaluation value.

[0067] Extract the upstream and downstream device sets of each energy-consuming device from the device topology and calculate the number of energy-dependent paths. Then, weight and sum the number of energy-dependent paths with the device efficiency evaluation value to obtain the device importance. Based on the device importance, identify key dependent devices and calculate the coupling dependency based on the connection strength of the key dependent devices in the device topology.

[0068] When collecting operating parameters and energy consumption parameters from various energy-consuming devices in an industrial system, operating parameters typically include real-time data such as equipment speed, temperature, pressure, flow rate, current, and voltage. Energy consumption parameters include power consumption, energy input, energy output, and energy loss for each device. The collected operating parameters are segmented according to a preset time window. The length of the time window is determined based on the typical operating cycle of the industrial system. For example, for a production line with shift switching patterns, the time window can be set to the duration of a complete shift to ensure that the segmented data covers the entire load change cycle. A frequency domain transformation is performed on the operating parameter sequence within each time window. Specifically, a Discrete Fourier Transform (DFT) is used to convert the time-domain signal into spectral components. The dominant frequency components and their corresponding amplitude distributions are extracted from these spectral components. The extraction of dominant frequency components is based on amplitude ranking, selecting the frequency components with the largest amplitudes as the dominant components. The amplitude distribution records the energy percentage at each dominant frequency. The dominant frequency components and amplitude distributions extracted from each device within each time window are combined according to a predetermined format to form a dynamic load feature vector that reflects the load fluctuation pattern of the device in different time periods. This vector contains both periodic information in the frequency dimension and intensity information in the amplitude dimension, thus fully characterizing the dynamic characteristics of the device operation.

[0069] Based on the energy input and output of each device in the energy consumption parameters, identify the energy transfer relationships between the devices. If the device Energy output and equipment If there is a significant numerical correlation between the energy input and the energy output, then it is determined that the energy is supplied by the equipment. Transfer to device The transmission direction is from point to Each energy-consuming device in an industrial system is abstracted as a node in a graph, with a total of [number missing] nodes. Each node identifies a direction of energy transfer, which is then abstracted as a directed edge. The starting point of each edge is the energy source device, and the ending point is the energy receiver device. Based on the definitions of nodes and directed edges, a device topology graph is constructed. , The set of nodes and edges in the diagram fully describes the energy flow network between various devices in an industrial system, i.e., the device topology. The accuracy of the device topology directly affects the reliability of subsequent efficiency assessments and path analysis. Therefore, during the construction process, it is necessary to perform threshold filtering on the energy transfer relationships to eliminate minor correlations caused by measurement noise and retain only energy transfer paths with actual physical significance.

[0070] The dynamic load characteristic vectors of each energy-consuming device are arranged in order of device number to form a load characteristic matrix. , where the matrix's first The row corresponds to the first The dynamic load characteristic vector of the equipment. Construct a topological adjacency matrix based on device topology relationships. , for The directed weight matrix, where elements Indicates equipment To the equipment The intensity of energy transfer; if there is no direct energy transfer relationship between the two, this element takes a value of zero. (For the load characteristic matrix) With topological adjacency matrix Perform matrix multiplication to obtain the neighborhood propagation feature matrix. The calculation formula is: , The The line represents the first The overall load propagation characteristics of a device after receiving energy from all its upstream devices. Calculate the neighborhood propagation characteristic matrix. With load characteristic matrix The row-by-row deviation is calculated using the Euclidean distance metric of each row vector to obtain the original deviation value for each device. The original deviation values ​​of all equipment are subjected to maximum-min normalization. The normalized result is the equipment efficiency evaluation value. , The larger the value, the greater the difference between the actual load characteristics of the equipment and its neighborhood propagation characteristics, which means that the equipment has a large efficiency loss in energy utilization and has high potential for energy-saving renovation.

[0071] Extract the upstream and downstream device sets for each device from the device topology. The upstream device set contains all devices that directly supply energy to that device, and the downstream device set contains all devices that directly receive energy from that device. For the first... For each device, count the number of all directed energy-dependent paths with that device as the endpoint, denoted as . The path statistics cover both direct connection paths and indirect dependent paths via intermediate nodes. The search depth of the paths is pre-set based on the size of the topology graph to control computational complexity. The number of energy-dependent paths is also considered. Equipment efficiency assessment value Perform a weighted summation to obtain the equipment importance. The calculation formula is: ,in and These are the path quantity weighting coefficient and the efficiency evaluation weighting coefficient, respectively, with a sum of 1. The specific values ​​are configured based on the industrial system's emphasis on energy-dependent complexity and equipment efficiency losses. Equipment Importance This comprehensively reflects the structural position of the equipment in the energy transfer network and its own efficiency status, with equipment of higher importance being identified as critical dependent equipment.

[0072] For the identified critical dependent devices, in the device topology diagram Extract the connection edges between the critical dependent device and its adjacent nodes, and calculate the connection strength of each edge. The connection strength is determined by the weight of the corresponding directed edge and the degree of bidirectional association of the edge. If a critical dependent device serves as the energy source for multiple downstream devices and also depends on the energy supply of multiple upstream devices, its connection strength is high. Summarize and calculate the connection edge strengths of all critical dependent devices to obtain the coupling dependency of that device. Coupling dependency The coupling degree between the device and the overall energy transfer network was quantified. A higher coupling dependency indicates that more related devices need to be considered simultaneously when modifying the device, and the higher the requirements for collaborative design of the modification scheme. The calculation results of the coupling dependency will serve as an important weighting basis in the subsequent identification of key energy consumption paths and analysis of energy consumption impact transmission chains, ensuring that the modification scheme can accurately grasp the mutual constraints between various devices in the industrial system, thereby generating a practically feasible energy-saving modification configuration.

[0073] In one alternative implementation,

[0074] Based on the coupling dependency, key energy consumption paths are identified, and propagation analysis is performed on the equipment efficiency evaluation values ​​to obtain the energy consumption impact transmission chain. Adjustability is determined based on the equipment efficiency evaluation values ​​of each node in the energy consumption impact transmission chain to obtain the modification potential value, including:

[0075] The connection paths between each energy-consuming device are extracted from the device topology. The cumulative coupling dependency value of each node on each connection path is calculated. Based on the cumulative coupling dependency value, all connection paths are sorted and filtered to obtain the key energy-consuming paths.

[0076] The equipment efficiency evaluation values ​​are arranged in the order of nodes in the key energy consumption path to form an initial efficiency vector. A path propagation matrix is ​​constructed based on the equipment topology relationship, and a first-order propagation efficiency vector is calculated by combining the initial efficiency vector with the path propagation matrix. An iterative matrix multiplication operation is performed on the first-order propagation efficiency vector and the path propagation matrix, and the results are accumulated to obtain a multi-order propagation efficiency vector. An energy consumption influence transmission chain is constructed based on the cumulative efficiency propagation value of each node in the multi-order propagation efficiency vector.

[0077] The efficiency sensitivity is obtained by extracting the equipment efficiency assessment value of each node from the energy consumption impact transmission chain and the corresponding efficiency propagation cumulative value in the multi-order propagation efficiency vector, calculating the ratio between the equipment efficiency assessment value and the efficiency propagation cumulative value, extracting the load fluctuation range of the corresponding equipment of each node from the dynamic load characteristics and calculating the adjustment response degree in combination with the efficiency sensitivity, and determining the adjustability based on the adjustment response degree and generating the transformation potential value.

[0078] After obtaining the coupling dependency and equipment efficiency assessment values ​​of each energy-consuming device, it is necessary to further identify the key energy consumption paths that have the most significant impact on the overall energy consumption in the industrial system, establish the energy consumption impact transmission chain through propagation analysis, and finally quantify the adjustability of each node to generate transformation potential values.

[0079] Extract all connection paths between energy-consuming devices from the device topology. A connection path is a complete directed sequence of paths in the device topology graph, starting from any initial device node, passing through several intermediate nodes, and reaching the final device node. For each connection path, sequentially read the coupling dependency of each node along the path direction, and sum the coupling dependencies of all nodes on the path to obtain the cumulative coupling dependency value for that path. ,in Number the path index. The calculation method is based on the path Coupling dependency of each node Summation, i.e. ,in Indicates the first The set of nodes contained in each path. After calculating the cumulative value of all connected paths, according to... All paths are sorted in descending order of energy consumption, and the paths with cumulative values ​​exceeding a preset threshold or ranking highly are selected as critical energy consumption paths. These paths represent the most concentrated energy flow and the most tightly coupled transmission channels between equipment in the industrial system, and are the core objects of subsequent energy-saving retrofit analysis.

[0080] After identifying the critical energy consumption path, the equipment efficiency assessment values ​​at each node are... An initial efficiency vector is constructed by arranging the nodes in the critical energy consumption path in sequence. Each component of the vector corresponds to the device efficiency evaluation value of a node on the path, and a path propagation matrix is ​​constructed based on the device topology. This matrix reflects the energy transfer weight relationship between adjacent nodes on the critical energy consumption path. Matrix elements Represents a node For nodes The efficiency propagation weights are determined by the ratio of energy flow between two nodes and the connection strength. For pairs of nodes that are not directly connected on the path, the corresponding matrix elements are zero. First-order propagation efficiency vector. By propagating the path matrix With the initial efficiency vector The result is obtained by performing matrix multiplication, i.e. This vector describes the distribution of the impact of each node's efficiency evaluation value at various locations along the path after one propagation.

[0081] To capture the multi-stage propagation effect of efficiency along the energy consumption path, an iterative matrix multiplication operation is performed on the first-order propagation efficiency vector and the path propagation matrix. Specifically, let the first-order propagation efficiency vector be multiplied by the first-order propagation matrix. The propagation efficiency vector is ,in The propagation order is determined by a value that starts from 1 and gradually increases to the preset maximum propagation order. The components of each order of propagation efficiency vector are summed to obtain the multi-order propagation efficiency vector. , its first Each component The calculation method is as follows ,in Indicates the first In the propagation efficiency vector of the first order The component values ​​corresponding to each node. That is, a node The cumulative efficiency propagation value comprehensively reflects the intensity of the multi-level propagation influence of this node along the entire energy consumption path. Based on the cumulative efficiency propagation value of each node... An energy consumption impact transmission chain is constructed using nodes as the basic unit. Each node in the transmission chain not only carries its own equipment efficiency assessment value, but also records the comprehensive impact it has on other nodes during multi-stage propagation, thus forming an ordered chain structure that can quantitatively describe the energy consumption impact transmission relationship.

[0082] After establishing the energy consumption impact transmission chain, the adjustability of each node is assessed to generate retrofit potential values. Equipment efficiency assessment values ​​for each node are then extracted from the energy consumption impact transmission chain. With the corresponding efficiency propagation cumulative value Calculate the ratio between the two to obtain the efficiency sensitivity of the node. ,Right now Efficiency sensitivity This reflects the cumulative impact that changes in unit equipment efficiency can have on the entire energy transmission chain: The larger the value, the greater the cascading effect. This indicates that even a small change in the efficiency of the node device will have a significant cascading effect on the transmission chain, thus having a higher value for energy-saving retrofits.

[0083] The load fluctuation range of the corresponding equipment at each node during the operating cycle is extracted from the dynamic load characteristics and denoted as . This represents the difference between the maximum and minimum loads exhibited by the device after time-domain decomposition. Load fluctuation range. This reflects the adjustable range of the equipment during actual operation; the wider the fluctuation range, the greater the load adjustment leeway of the equipment. Combined with efficiency sensitivity... With load fluctuation range Calculate the regulation response of each node. The calculation method is as follows This indicator comprehensively measures the influence of a node on efficiency propagation and its adjustability in actual operation, and is the core basis for judging whether equipment is ready for modification.

[0084] Based on regulation response Perform adjustability determination: when Exceeding the preset adjustability threshold When the node is identified as having potential for modification, it is determined that the corresponding equipment has such potential. If the value is below the threshold, the device is deemed to have limited potential for modification in its current operating state and is not included in the priority modification scope. For nodes determined to have modification potential, their responsiveness is adjusted. The normalized value is used as the transformation potential value of this node. The normalization method is to... Divide by the maximum adjustment response among all nodes with modification potential ,Right now This ensures that the potential value for modification falls within the range of 0 to 1, facilitating subsequent horizontal comparison and ranking of the modification priorities for different equipment. Modification Potential Value The closer the value is to 1, the higher the priority of the equipment in energy-saving renovation, and it should be given priority consideration in the subsequent reverse tracing and target equipment selection stages.

[0085] In one alternative implementation,

[0086] Based on the transformation potential value, the energy consumption source nodes are identified by reverse tracing along the energy consumption impact transmission chain. The fluctuation characteristics are determined based on the dynamic load characteristics corresponding to the energy consumption source nodes and matched with the pre-configured equipment adjustment capabilities to calculate the transformation adaptability. Based on the transformation adaptability, target transformation equipment is selected and determined, including:

[0087] The potential values ​​of the transformation are arranged into potential vectors according to the nodes in the energy consumption impact transmission chain. A back propagation matrix is ​​constructed based on the equipment topology. The importance weight of the nodes corresponding to the potential vector is determined based on the back attention mechanism. The weighted back tracing vector is obtained by combining the back propagation matrix. The energy contribution gradient is extracted from the weighted back tracing vector and the node corresponding to the gradient vanishing position is identified as the energy consumption source node.

[0088] The load time series data of the equipment corresponding to the energy source node is extracted from the dynamic load characteristics. Wavelet packet decomposition is performed on the load time series data to obtain fluctuation components of different frequency bands. The dominant frequency band is identified from the fluctuation components and instantaneous energy is extracted. An energy envelope curve is constructed based on the instantaneous energy and the peak distribution pattern is extracted to obtain the fluctuation characteristics.

[0089] The system acquires the pre-configured equipment adjustment capability and extracts the adjustment response bandwidth and energy buffer capacity. It extracts the frequency domain response requirement and instantaneous power surge from the fluctuation characteristics. It performs matching calculations between the frequency domain response requirement and the adjustment response bandwidth to obtain the frequency adaptation coefficient. It performs matching calculations between the instantaneous power surge and the energy buffer capacity to obtain the capacity adaptation coefficient. It then combines the frequency adaptation coefficient to solve for the modification adaptation degree. Based on the modification adaptation degree, it screens and determines the target modification equipment.

[0090] After obtaining the modification potential value of each node, it is necessary to trace back along the energy consumption impact transmission chain to identify the source node that truly drives energy loss. The modification potential values ​​of each node are then arranged in order of node sequence in the energy consumption impact transmission chain to form a potential vector. , of which Each component corresponds to a node Potential for transformation Based on the established device topology, the forward propagation matrix is ​​transposed to obtain the backward propagation matrix. The matrix describes the reverse mapping relationship of energy consumption impact from downstream nodes to upstream nodes in the transmission chain, so that energy contribution can be traced step by step from the end effect node to the source.

[0091] Introducing a reverse attention mechanism for potential vectors The importance of each node in the transmission chain is evaluated using a weighted average. For each node in the transmission chain... Calculate their attention score The score takes into account the node's hierarchical depth in the transmission chain, the relative magnitude of its transformation potential, and the strength of its energy consumption correlation with neighboring nodes. By analyzing the nodes The transformation potential value is obtained by normalizing the product of its value and the sum of the weights of all outgoing edges in the topology. The normalization uses a softmax transformation to ensure that the sum of the attention scores of all nodes is 1. The attention scores are then arranged into a diagonal matrix. , and the backpropagation matrix Perform joint operations to obtain the weighted reverse tracing vector. The calculation method is as follows ,in The Each component Reflects the nodes The intensity of the overall energy contribution in reverse tracing.

[0092] Tracing back the vector from the weighted reverse direction Extract the energy contribution gradient. Along the node sequence of the transmission chain... Perform a difference operation to obtain the gradient value sequence between adjacent nodes. ,in Number the node's position in the propagation chain. When the gradient value The absolute value is lower than the preset gradient vanishing threshold. When a gradient vanishing occurs at a given location, it is considered that the energy contribution no longer propagates significantly upstream. After identifying all gradient vanishing locations, nodes upstream of these locations that still maintain a large energy contribution gradient are identified as energy source nodes. These nodes are the fundamental driving points causing downstream energy loss and have the highest priority for modification.

[0093] After identifying the energy consumption source nodes, load time-series data of the corresponding equipment is extracted from the dynamic load characteristics. Load time-series data is typically obtained at fixed sampling intervals. Record the actual power or flow data of the device within a historical time window to form a time series. ,in The total number of sampling points. For the first Each sampling time point. Wavelet packet decomposition is performed on this time-series data, selecting the db4 wavelet basis function, and the number of decomposition levels is set to [value missing]. Layers, thereby decomposing the original load signal into The system consists of several fluctuation components at different frequency bands, each corresponding to load fluctuation information within a specific frequency range. Low-frequency components reflect slow, trend-based changes in the equipment, high-frequency components reflect rapid, transient impact behaviors, and mid-frequency components correspond to regular fluctuations caused by periodic process cycles.

[0094] Identify the dominant frequency band from the fluctuation components of each frequency band. Calculate the energy percentage of each frequency band component. ,in For frequency band numbering, This is equal to the sum of the squares of the amplitudes of all sampled points of the component in that frequency band divided by the total energy of all frequency bands. The frequency bands with the largest energy percentage are determined as the dominant frequency bands; typically, the set of frequency bands with a cumulative energy percentage exceeding 85% is selected as the dominant frequency band set. The instantaneous energy of the dominant frequency band component is extracted using the Hilbert transform. Instantaneous energy reflects the local energy intensity of the load signal at various times. An energy envelope curve is constructed based on the instantaneous energy sequence, and a smooth energy envelope is obtained using an upper envelope extraction algorithm (connecting all local maxima points and performing cubic spline interpolation). By statistically analyzing the distribution of peak occurrence time intervals, peak amplitudes, and peak durations from the energy envelope curve, a peak distribution pattern is formed, ultimately providing a complete description of the fluctuation characteristics of the equipment corresponding to the energy source node.

[0095] Obtain pre-configured device tuning capability parameters, including adjusting response bandwidth. and energy buffer capacity Adjusting the response bandwidth This indicates the upper limit of the load fluctuation frequency that the equipment can effectively respond to, measured in Hz, reflecting the equipment's ability to follow rapidly changing loads; energy buffer capacity. This represents the maximum amount of energy that the equipment can absorb or release in a short period of time, measured in kWh, reflecting the equipment's ability to withstand instantaneous power surges. Frequency domain response requirements are extracted from fluctuation characteristics. and instantaneous power surge ,in Take the highest frequency value of the dominant frequency band. Take the 95th percentile peak power from the peak distribution of the energy envelope curve.

[0096] The frequency domain response requirement is matched with the adjustment response bandwidth to obtain the frequency adaptation coefficient. The calculation method is as follows When the adjustment response bandwidth is greater than or equal to the frequency domain response requirement, the frequency adaptation coefficient is 1, indicating that the equipment fully meets the modification requirements in the frequency dimension. When the adjustment response bandwidth is less than the frequency domain response requirement, the frequency adaptation coefficient is less than 1, reflecting the degree of inadequacy in frequency tracking capability. The capacity adaptation coefficient is obtained by matching the instantaneous power surge with the energy buffer capacity. The calculation method is as follows ,in The average duration of peak power. This indicates the amount of energy buffer required for a single impact event.

[0097] Frequency adaptation coefficient With capacity adaptation factor Weighted fusion is performed to obtain the final modification adaptability. The calculation formula is: ,in and These are the fusion weights for the frequency adaptation coefficient and the capacity adaptation coefficient, respectively, satisfying... The weight values ​​are pre-configured based on the industrial system's emphasis on frequency response and energy buffering requirements. Adaptability to modification. The value ranges from 0 to 1. The larger the value, the better the equipment's adjustment capability matches the fluctuation characteristics, and the higher the reliability of achieving energy-saving effects after the modification.

[0098] Based on the adaptability of the modification Candidate devices corresponding to energy consumption source nodes are screened, and a suitability screening threshold is set. The modification will have a higher compatibility than Equipment with a suitable adaptability to the target modification set was included. For equipment with a modification adaptability below the threshold, although its modification potential was high, forced modification could lead to operational instability due to a significant gap between its regulation capacity and actual fluctuation requirements. Therefore, this equipment was marked as requiring an upgrade to its regulation capacity before modification and was not included in the current batch of target modification equipment. The final determined target modification set comprehensively reflects the dual constraints of modification potential and feasibility, providing a reliable range of equipment for subsequent optimization of energy-saving operating parameters.

[0099] Figure 2 This is a flowchart illustrating the energy consumption source tracing process of the intelligent generation method for industrial energy-saving renovation schemes according to an embodiment of the present invention.

[0100] In one alternative implementation,

[0101] The transformation potential values ​​are arranged into potential vectors according to the nodes in the energy consumption impact transmission chain. A backpropagation matrix is ​​constructed based on the equipment topology. The node importance weights corresponding to the potential vectors are determined based on the back attention mechanism, and the weighted back tracing vector is obtained by combining the backpropagation matrix.

[0102] Extract the transformation potential value of each node from the energy consumption impact transmission chain and arrange them into a potential vector according to the node order. Extract the upstream dependency relationship between each energy-consuming device from the device topology relationship. Reverse the energy transmission direction in the upstream dependency relationship and construct a reverse connection graph. Calculate the node reverse propagation coefficient based on the in-degree and out-degree of each node in the reverse connection graph and fill it into a reverse propagation matrix according to the node position.

[0103] The potential vector is used as the query vector, and the cumulative efficiency propagation value of each node is extracted from the energy consumption influence transmission chain as the key vector. The query vector and the key vector are multiplied by a dot product and a scaling transformation is applied to obtain the attention score. The attention score is normalized to obtain the attention weight distribution. The potential vector is weighted and summed based on the attention weight distribution to obtain the context-aware potential vector. The context-aware potential vector and the potential vector are residually connected and adjusted by a gating unit to obtain the node importance weight.

[0104] The node importance weights are converted into diagonal matrices and multiplied with the backpropagation matrix to obtain a weighted backpropagation matrix. The potential vector is then multiplied with the weighted backpropagation matrix using iterative matrix multiplication, and the results are accumulated to obtain a weighted backtracking vector.

[0105] Extract the transformation potential value of each node from the energy consumption impact transmission chain, and assemble them into a potential vector according to the order of the nodes in the transmission chain. Each component of this vector corresponds to the transformation potential value of a node on the transmission chain. When extracting upstream dependencies between energy-consuming devices in the device topology, it is necessary to identify the energy source direction of each device, i.e., which devices supply energy to the current device and which devices receive the output energy from the current device. The energy transfer directions in these upstream dependencies are all reversed. Originally, directed edges pointing from upstream nodes to downstream nodes become directed from downstream nodes to upstream nodes after reversal, thus constructing a reverse connection graph. The reverse connection graph retains the structural information of the original topology, but all edges are oriented in the opposite direction to the original energy flow, allowing subsequent reverse tracing to perform source analysis in the opposite direction of energy consumption propagation.

[0106] In the reverse connectivity graph, the in-degree of each node is calculated separately. With out The in-degree represents the number of edges pointing to a node in the reverse graph, and the out-degree represents the number of edges originating from that node pointing to other nodes. The backpropagation coefficient of a node is calculated based on its in-degree and out-degree. The calculation method is as follows ,in To prevent extremely small positive numbers with a denominator of zero, the backpropagation coefficient reflects the degree to which a node acts as an energy convergence point in the backpropagation graph. Nodes with larger in-degrees are more likely to become the convergence point of multiple paths during backpropagation, and their backpropagation coefficients are also larger. The backpropagation coefficients of each node are filled into a backpropagation matrix according to the node's position. Matrix elements Represents a node In the reverse graph, the nodes The coefficient of propagation energy consumption impact, if there are no nodes in the reverse graph. To the node The edge, then If a corresponding edge exists, then This allows the backpropagation matrix to simultaneously encode the back-connection structure and the propagation strength of each node.

[0107] Potential vector As a query vector, the cumulative efficiency propagation value of each node is extracted from the energy consumption impact propagation chain. And assembled into a key vector ,in The Each component is a node. Efficiency propagation cumulative value For query vectors With key vector Perform a dot product operation to obtain the original attention score vector, whose _i_th ... Each component is This reflects the correlation between the distribution of transformation potential and the cumulative value of efficiency propagation at each node. To prevent the dot product result from becoming too large and causing gradient instability due to its large vector dimension, a scaling transformation is applied to the dot product result, dividing each component by... ,in Given the dimension of the key vector, we obtain the scaled attention score. The attention score vector is then normalized using softmax to obtain the attention weight distribution. ,satisfy And each A node with a higher attention weight indicates a stronger match between its cumulative efficiency propagation value and the overall transformation potential distribution.

[0108] Based on attention weight distribution on potential vector Perform a weighted summation to obtain the context-aware potential vector. The calculation method is as follows ,in For the first The position basis vectors of each node. The context-aware potential vector, based on the original potential distribution, incorporates the modulation effect of the efficiency propagation characteristics of each node on the transformation potential, further enhancing the transformation potential of nodes with high efficiency propagation. The context-aware potential vector... With the original potential vector Perform residual joins to obtain the residual fusion vector. The introduction of residual connections effectively prevents the attention mechanism from excessively suppressing the original potential information, preserving the transformation potential features of each node. Subsequently, the residual fusion vector is adjusted by a gating unit, which uses a sigmoid activation function to generate the gating coefficient vector. , its first Each component ,in For gating weights, As a gated bias, the gate coefficient controls the proportion of each component retained in the residual fusion vector, ultimately yielding the node importance weight vector. Each of its components That is, a node The node importance weighting mechanism allows the node importance weighting to adaptively suppress nodes with high noise or low modification value, thereby improving the accuracy of reverse tracing.

[0109] Node importance weight vector Convert to a diagonal matrix , matrix number Line 1 The elements of the column are All off-diagonal elements are zero. Transform the diagonal matrix... With backpropagation matrix Perform matrix multiplication to obtain the weighted backpropagation matrix. The weighted backpropagation matrix modulates the propagation coefficient of each row with the corresponding node importance weight, based on the backpropagation matrix. This makes nodes with high importance gain stronger propagation influence in back tracing, while the propagation coefficients of nodes with low importance are compressed accordingly.

[0110] For potential vector with weighted backpropagation matrix Perform iterative matrix multiplication, let the number of iterations be . , No. The intermediate result of the second iteration is ,in Represents the weighted backpropagation matrix The power of 1. Each iteration propagates the current potential distribution one step upstream along the backpropagation path. After multiple iterations, the indirect contribution of distant upstream nodes to the overall transformation potential can also be effectively captured. The intermediate results of each iteration are accumulated and summed to obtain the weighted backpropagation vector. , its first Each component Comprehensive reflection of nodes The cumulative contribution to the overall transformation potential across all stages of dissemination. The larger the value, the more likely that node is the source of the energy consumption problem in the reverse tracing process. Further analysis will be conducted based on the values ​​of each node. The system identifies and filters energy consumption source nodes, providing a basis for precise positioning of the renovation plan. (Number of iterations) The setting needs to be balanced with the depth of the transmission chain and computing resources. Usually, half of the total number of nodes in the transmission chain is taken as the upper limit to ensure that the tracing depth covers the main energy consumption propagation path, while avoiding the loss of distinguishability due to excessive exponentiation operations causing matrix elements to approach zero.

[0111] In one alternative implementation,

[0112] Based on the dynamic load characteristics and preset process constraints, an operating state constraint space corresponding to the target modified equipment is constructed, and the energy consumption parameters are optimized and solved to obtain energy-saving operating parameters, including:

[0113] Based on the dynamic load characteristics, the load change rate is determined. The process flow stability boundary and capacity guarantee threshold are extracted from the preset process constraints. The dynamic constraint boundary is determined based on the load change rate and the process flow stability boundary, and the static constraint boundary is determined based on the capacity guarantee threshold. The intersection operation of the dynamic constraint boundary and the static constraint boundary is performed to obtain the feasible operating domain.

[0114] Based on the feasible operating domain, a set of candidate operating points is divided and the energy input to output efficiency ratio of each candidate operating point is calculated to obtain an initial efficiency score. Downstream related devices are extracted from the energy consumption influence transmission chain and the energy consumption transmission influence of each candidate operating point on the downstream related devices is calculated. The energy consumption transmission influence and the initial efficiency score are weighted and summed to obtain a global efficiency score.

[0115] Based on the global efficiency score, a subset of high-efficiency operating points is obtained. The operating points in the subset of high-efficiency operating points are time-series aligned with the dynamic load characteristics, and the time offset is calculated to obtain the response delay coefficient. The global efficiency score and the response delay coefficient are normalized and weighted to obtain a comprehensive evaluation index. Based on the comprehensive evaluation index, the optimal operating point is selected and the corresponding control parameter combination is extracted to obtain energy-saving operating parameters.

[0116] After identifying the target equipment for modification, it is necessary to construct an operational state constraint space for the equipment and optimize the energy consumption parameters within the constraint space to obtain energy-saving operating parameters. The core of this process lies in organically integrating dynamic load behavior with the rigid boundaries of the process to form a feasible operating region that meets both production stability requirements and has energy-saving potential.

[0117] Extracting the load change rate from dynamic load characteristics is fundamental to constructing dynamic constraint boundaries. Differential processing is performed on the load time-series data of the target modified equipment, calculating the ratio of load increment to time interval between adjacent sampling moments to obtain the instantaneous load change rate sequence at each moment. This sequence is averaged using a sliding window to eliminate short-term noise interference, resulting in a smoothed load change rate curve. The load change rate reflects the dynamic response requirements of the equipment under actual operating conditions. An excessively large change rate means the equipment needs to complete a significant power adjustment within a short period, placing high pressure on the equipment's mechanical structure and control system. A process stability boundary is extracted from a pre-defined library of process constraints. This boundary is typically represented by the upper limit of the allowable rate of change of equipment operating parameters, such as the compressor speed change rate not exceeding a certain set value, or the heat exchanger inlet and outlet temperature difference change rate not exceeding the process-permitted range. The smoothed load change rate is compared with the process stability boundary moment by moment, determining the dynamically adjustable range of the equipment operating parameters at each moment, thus forming a time-varying dynamic constraint boundary.

[0118] The static constraint boundary is determined from the capacity guarantee threshold. The capacity guarantee threshold specifies the minimum output capacity that equipment must maintain per unit time, such as the minimum flow requirement for a pump set or the minimum torque output guarantee for a motor. The capacity guarantee threshold is converted into a lower bound constraint in the equipment's operating parameter space, and simultaneously, combined with the equipment's nameplate rated parameters, an upper bound constraint for the operating parameters is determined. Both together constitute the static constraint boundary. The static constraint boundary does not change over time and represents a hard limit on the equipment's operating parameters; no candidate operating point may exceed this boundary range.

[0119] The feasible operating region is obtained by performing an intersection operation on the dynamic and static constraint boundaries. The feasible operating region is the common feasible interval of the two types of constraints, and is represented in the parameter space as a multidimensional feasible region that evolves over time. At each time step, the boundary of the feasible operating region is jointly determined by the more stringent side of the dynamic and static constraint boundaries. If the dynamic constraint boundary is more stringent than the static constraint boundary at a certain time, the dynamic constraint boundary prevails; otherwise, the static constraint boundary prevails. After the intersection operation, all operating points within the feasible operating region simultaneously satisfy the process stability requirements and capacity assurance requirements, forming a valid search space for subsequent optimization solutions.

[0120] Within the feasible operating domain, candidate operating points are divided into sets using either uniform grid sampling or adaptive sampling based on historical operating condition density. For each candidate operating point, its corresponding energy input and output efficiency are extracted, and the ratio of these two is calculated to obtain an initial efficiency score. A higher initial efficiency score indicates that the operating point can achieve a higher effective output per unit energy input, possessing superior energy-saving performance. Let the initial efficiency score be... ,in Number the candidate run points. It is obtained by dividing the output efficiency at that operating point by the corresponding energy input.

[0121] Considering only the efficiency of the target modified equipment is insufficient to fully assess the energy-saving value of an operating point; its energy transfer impact on downstream related equipment must also be considered. A set of downstream related equipment directly or indirectly connected to the target modified equipment is extracted from the energy transfer chain. For each candidate operating point, based on the equipment topology and energy transfer model, the impact of the energy parameters (such as flow rate, pressure, temperature, etc.) output by that operating point on the energy consumption of each downstream device after being transferred to them is calculated. ,in Number the candidate operating points. A negative impact value indicates that the operating point can reduce the energy consumption of downstream equipment, while a positive value indicates that it will increase the energy consumption burden of downstream equipment. The initial efficiency score and the energy consumption propagation impact value are weighted and summed to obtain the global efficiency score. The calculation method is as follows ,in For the ontology efficiency weight, The downstream impact weight is calculated, and the sum of the two is 1. The specific value is configured according to the emphasis of the process system on energy saving of the device itself and global energy saving. The global efficiency score comprehensively measures the energy saving contribution of the target modified equipment to the entire energy consumption impact transmission chain at a certain operating point, and is a more comprehensive operating point evaluation index.

[0122] All candidate operating points are sorted in descending order based on a global efficiency score, and a subset of operating points with scores ranking in the top predefined proportion are selected. For each operating point in the subset of efficient operating points, a time-series alignment analysis is performed with dynamic load characteristics. The purpose of time-series alignment is to evaluate the response timeliness of the operating point, i.e., whether the control parameter combination of the operating point can complete the adjustment response within a reasonable time window after the load change occurs during actual execution. Specifically, the cross-correlation calculation is performed between the control parameter change sequence corresponding to each operating point and the dynamic load characteristic sequence to find the time offset that maximizes the correlation between the two. This offset is the response delay coefficient at that running point. The smaller the value, the more timely the control parameters at that operating point can respond to load changes, and the better the dynamic adaptability. An excessively large value indicates a significant lag in the adjustment of control parameters, which may lead to energy waste or process deviations during periods of rapid load change.

[0123] The global efficiency score and response delay coefficient are normalized to eliminate dimensional differences. The normalized global efficiency score is denoted as... The normalized response delay coefficient is denoted as its reciprocal. This is to ensure that a smaller delay corresponds to a greater evaluation contribution. A weighted sum of the two factors yields a comprehensive evaluation index. The calculation method is as follows ,in To incorporate weights into the efficiency score, To account for the delayed fusion weights, the sum of the two is 1. Comprehensive evaluation index. It also reflects the overall energy-saving benefits and dynamic response capability of the operating point, and serves as the basis for ultimately selecting the optimal operating point.

[0124] The operating point with the highest comprehensive evaluation index is selected as the optimal operating point. The corresponding combination of control parameters is extracted from this operating point, including specific control quantities such as the target equipment's speed setpoint, power limit, pressure setpoint, and flow regulation command, forming a set of energy-saving operating parameters. These energy-saving operating parameters not only satisfy all constraints within the feasible operating domain but also achieve an optimal balance between global energy-saving benefits and dynamic response capabilities. They can be directly used for subsequent forward propagation calculations and collaborative modification configuration generation.

[0125] In one alternative implementation,

[0126] The energy-saving operating parameters are forward-propagated along the energy consumption impact transmission chain to calculate the parameter adjustment amount for each associated device. Based on the parameter adjustment amount and the device topology relationship, a collaborative transformation configuration is generated and a transformation plan is output, including:

[0127] Extract the control parameter combination and energy input change value of the target modified equipment from the energy-saving operation parameters, extract the propagation direction and energy transfer coefficient in the energy consumption influence transmission chain, propagate the energy input change value to downstream related equipment layer by layer and accumulate to obtain the energy consumption change of each related equipment, and perform reverse calculation based on the energy consumption change and the equipment efficiency evaluation value of each related equipment to obtain the parameter adjustment amount of each related equipment.

[0128] The parallel and series relationships between devices are extracted from the device topology. Based on the parallel relationship, the parameter adjustment amount in the parallel device group is load balanced and distributed to obtain the balanced adjustment amount. Based on the series relationship, the parameter adjustment amount in the series device chain is time-series coordinated and corrected to obtain the coordinated adjustment amount.

[0129] The control parameters of the target equipment to be modified, along with the equalization adjustment amount and the coordination adjustment amount, are combined according to the equipment hierarchy in the equipment topology to construct a collaborative modification configuration and output a modification scheme.

[0130] After determining the energy-saving operating parameters of the target equipment, it is necessary to spread the impact of these parameters downstream along the energy consumption impact transmission chain, calculate the parameter adjustment amount required for each related equipment, and finally form a coordinated transformation configuration covering the whole.

[0131] Extracting the control parameter combination and energy input change value of the target modified equipment from the energy-saving operation parameters is the starting point of forward propagation. The control parameter combination includes controllable quantities such as the speed setpoint, load rate, and switching sequence of the target modified equipment under optimized operating conditions, while the energy input change value represents the difference in energy consumption of the equipment under energy-saving operation conditions compared to the original operating conditions. Let the energy input value of the target modified equipment before energy-saving optimization be... The optimized energy input is The change in energy input is then... This value is usually negative, indicating reduced energy consumption. As the starting node of the forward propagation driving force, the energy consumption effect is injected into the transmission chain.

[0132] In the energy consumption impact propagation chain, each directed edge carries the propagation direction and energy transfer coefficient. Energy transfer coefficient Indicates upstream node What percentage of the energy consumption changes will be passed on to downstream nodes? Its value is determined by the physical coupling relationship and process flow between the two devices, and is usually obtained through regression fitting of historical operating data during the device topology relationship construction stage. During layer-by-layer propagation, the downstream device nodes of the first layer in the transmission chain... The received energy consumption change is For deeper downstream nodes Its energy consumption change is the cumulative amount of input from all upstream nodes in the current layer, i.e. ,in Represents a node The set of all direct predecessor nodes in the transmission chain. This process is iterated layer by layer until the end of the transmission chain, obtaining the energy consumption changes of all associated devices. .

[0133] After obtaining the energy consumption changes of each associated device, the equipment efficiency assessment values ​​are used for inverse calculation to obtain the operating parameters that need to be adjusted for each associated device. For the associated devices... In terms of its equipment efficiency assessment value This describes the effective output corresponding to the unit energy consumption of the device under the current operating conditions. Therefore, when the energy consumption change is... At that time, the effective output change of the device is Furthermore, based on the equipment's rated operating condition curve or a pre-calibrated parameter-output mapping relationship, This is reverse-mapped to specific controllable parameter adjustments, such as the speed adjustment of pumps, the pressure ratio adjustment of compressors, and the flow rate adjustment of heat exchangers, denoted as... This inverse mapping relationship can be implemented using piecewise linear interpolation or a polynomial fitting model based on equipment nameplate data to ensure that the calculation accuracy of parameter adjustments meets the requirements of engineering implementation.

[0134] Extracting parallel and series relationships from the equipment topology is a key step in generating a coordinated upgrade configuration. In a parallel equipment group, each device shares the same energy supply task. If their parameters are adjusted independently based on their individual energy consumption changes, uneven load distribution within the group may occur, leading to localized overloads or efficiency degradation. Therefore, load balancing is performed on the parallel equipment group: Assuming the parallel group has a total of... The equipment has a raw load rate of [number] units. ( The parameter adjustments obtained by each device after forward propagation are: Then the balance adjustment amount Under the constraint of total quantity conservation Under the premise of current efficiency evaluation value of each device Perform weighted allocation: The allocation method allows more efficient equipment to handle more adjustments, thereby improving the overall operating efficiency of the parallel group.

[0135] In a series of connected devices, the devices in the process flow are sequentially dependent. Parameter adjustments to upstream devices must be completed before those to downstream devices; otherwise, unmet input conditions may lead to process fluctuations or even shutdowns. The devices in the series chain are numbered according to their topological order as follows: Let the first The parameter adjustment amount of the series-connected equipment is Its response latency is (This represents the time required for the device to complete the parameter switching from receiving the adjustment command), then the coordinated adjustment amount. Numerically with Same, but with additional timing constraints: The actual execution time of the device satisfy This means that adjustments to each piece of equipment can only be initiated after adjustments to its directly upstream equipment have been completed and stabilized. The purpose of timing coordination correction is to minimize the overall adjustment cycle and reduce energy loss during transition periods while ensuring process continuity.

[0136] The collaborative modification configuration is constructed by combining the control parameters of the target modified equipment, the balancing adjustment amount of the parallel equipment group, and the coordination adjustment amount of the series equipment chain according to the equipment hierarchy in the equipment topology. The equipment hierarchy is divided according to the depth of the equipment in the energy consumption impact transmission chain, with the target modified equipment located at level 0, its directly downstream equipment located at level 1, and so on. The collaborative modification configuration is organized in a hierarchical structure, with each level containing a set of adjustment instructions for all equipment at that level. The adjustment instructions include the equipment identifier, parameter adjustment amount (balancing adjustment amount or coordination adjustment amount), execution timing constraints, and a description of the expected operating state after parameter adjustment.

[0137] The output of the retrofit plan is presented in the form of a structured report, which includes the energy-saving operating parameter settings for the target retrofit equipment, a list of parameter adjustments for each related device, a load balancing allocation scheme for parallel groups, a time-series coordination execution plan for series chains, and an estimated total system energy saving. The expected energy saving can be calculated by analyzing the energy consumption changes of each device. The summation yields the total energy saved by the entire system. ,in This refers to the collection of all downstream related equipment nodes. This modification plan can be directly used to guide field engineers in equipment parameter tuning and modification implementation, and can also be further imported into industrial control systems as a configuration file for automated execution.

[0138] A second aspect of this invention provides an intelligent generation system for industrial energy-saving retrofit schemes, comprising:

[0139] The coupling calculation unit is used to obtain the operating parameters and energy consumption parameters of each energy-consuming device in the industrial system, perform time-domain decomposition on the operating parameters to obtain dynamic load characteristics, construct the device topology relationship based on the energy consumption parameters, and perform correlation calculation with the dynamic load characteristics to obtain the device efficiency evaluation value and coupling dependency.

[0140] The propagation analysis unit is used to identify key energy consumption paths based on the coupling dependency and perform propagation analysis on the equipment efficiency evaluation value to obtain the energy consumption impact transmission chain. Based on the equipment efficiency evaluation value of each node in the energy consumption impact transmission chain, the unit performs adjustability determination to obtain the transformation potential value.

[0141] The source matching unit is used to identify energy consumption source nodes by tracing back along the energy consumption impact transmission chain based on the transformation potential value, determine the fluctuation characteristics based on the dynamic load characteristics corresponding to the energy consumption source nodes and match them with the pre-configured equipment adjustment capabilities to calculate the transformation adaptability, and screen and determine the target transformation equipment based on the transformation adaptability.

[0142] The scheme generation unit is used to construct the operating state constraint space corresponding to the target modification equipment based on the dynamic load characteristics and preset process constraints, optimize and solve the energy consumption parameters to obtain energy-saving operating parameters, forward propagate the energy-saving operating parameters along the energy consumption influence transmission chain to calculate the parameter adjustment amount of each associated equipment, generate a collaborative modification configuration based on the parameter adjustment amount and the equipment topology relationship, and output the modification scheme.

[0143] A third aspect of the present invention provides an electronic device, comprising:

[0144] A processor and a memory for storing processor-executable instructions, wherein the processor is configured to invoke instructions stored in the memory to perform the aforementioned method.

[0145] A fourth aspect of the present invention provides a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the aforementioned method.

[0146] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligently generating industrial energy-saving renovation schemes, characterized in that, include: The operating parameters and energy consumption parameters of each energy-consuming device in the industrial system are obtained. The operating parameters are decomposed in the time domain to obtain dynamic load characteristics. Based on the energy consumption parameters, the device topology relationship is constructed and correlated with the dynamic load characteristics to obtain the device efficiency evaluation value and coupling dependency. Based on the coupling dependency, key energy consumption paths are identified and propagation analysis is performed on the equipment efficiency evaluation values ​​to obtain the energy consumption impact transmission chain. Based on the equipment efficiency evaluation values ​​of each node in the energy consumption impact transmission chain, adjustability is determined to obtain the transformation potential value. Based on the transformation potential value, the energy consumption source node is identified by reverse tracing along the energy consumption impact transmission chain. The fluctuation characteristics are determined based on the dynamic load characteristics corresponding to the energy consumption source node and matched with the pre-configured equipment adjustment capabilities to obtain the transformation adaptability. The target transformation equipment is selected based on the transformation adaptability. Based on the dynamic load characteristics and preset process constraints, the operating state constraint space corresponding to the target modification equipment is constructed, and the energy consumption parameters are optimized and solved to obtain energy-saving operating parameters. The energy-saving operating parameters are forward propagated along the energy consumption influence transmission chain to calculate the parameter adjustment amount of each associated equipment. Based on the parameter adjustment amount and the equipment topology relationship, a collaborative modification configuration is generated and a modification scheme is output.

2. The method according to claim 1, characterized in that, The system acquires the operating parameters and energy consumption parameters of each energy-consuming device in the industrial system. It then performs time-domain decomposition on the operating parameters to obtain dynamic load characteristics. Based on the energy consumption parameters, it constructs device topology relationships and performs correlation calculations with the dynamic load characteristics to obtain device efficiency evaluation values ​​and coupling dependencies, including: The operating parameters and energy consumption parameters of each energy-consuming device in the industrial system are obtained. The operating parameters are segmented according to a preset time window and frequency domain transformation is performed to obtain spectral components. The dominant frequency components and amplitude distribution are extracted from the spectral components and combined to obtain dynamic load characteristics. Based on the energy input-output relationship in the energy consumption parameters, identify the energy transfer direction between each energy-consuming device, construct a device topology graph by taking each energy-consuming device as a node and the energy transfer direction as a directed edge, and determine the device topology relationship based on the device topology graph; The dynamic load characteristics are arranged into a load characteristic matrix according to the energy-consuming equipment. A topological adjacency matrix is ​​constructed based on the equipment topology relationship. The load characteristic matrix and the topological adjacency matrix are multiplied to obtain the neighborhood propagation characteristics. The deviation between the neighborhood propagation characteristics and the load characteristic matrix is ​​calculated and normalized to obtain the equipment efficiency evaluation value. Extract the upstream and downstream device sets of each energy-consuming device from the device topology and calculate the number of energy-dependent paths. Then, weight and sum the number of energy-dependent paths with the device efficiency evaluation value to obtain the device importance. Based on the device importance, identify key dependent devices and calculate the coupling dependency based on the connection strength of the key dependent devices in the device topology.

3. The method according to claim 1, characterized in that, Based on the coupling dependency, key energy consumption paths are identified, and propagation analysis is performed on the equipment efficiency evaluation values ​​to obtain the energy consumption impact transmission chain. Adjustability is determined based on the equipment efficiency evaluation values ​​of each node in the energy consumption impact transmission chain to obtain the modification potential value, including: The connection paths between each energy-consuming device are extracted from the device topology. The cumulative coupling dependency value of each node on each connection path is calculated. Based on the cumulative coupling dependency value, all connection paths are sorted and filtered to obtain the key energy-consuming paths. The equipment efficiency evaluation values ​​are arranged in the order of nodes in the key energy consumption path to form an initial efficiency vector. A path propagation matrix is ​​constructed based on the equipment topology relationship, and a first-order propagation efficiency vector is calculated by combining the initial efficiency vector with the path propagation matrix. An iterative matrix multiplication operation is performed on the first-order propagation efficiency vector and the path propagation matrix, and the results are accumulated to obtain a multi-order propagation efficiency vector. An energy consumption influence transmission chain is constructed based on the cumulative efficiency propagation value of each node in the multi-order propagation efficiency vector. The efficiency sensitivity is obtained by extracting the equipment efficiency assessment value of each node from the energy consumption impact transmission chain and the corresponding efficiency propagation cumulative value in the multi-order propagation efficiency vector, calculating the ratio between the equipment efficiency assessment value and the efficiency propagation cumulative value, extracting the load fluctuation range of the corresponding equipment of each node from the dynamic load characteristics and calculating the adjustment response degree in combination with the efficiency sensitivity, and determining the adjustability based on the adjustment response degree and generating the transformation potential value.

4. The method according to claim 1, characterized in that, Based on the transformation potential value, the energy consumption source nodes are identified by reverse tracing along the energy consumption impact transmission chain. The fluctuation characteristics are determined based on the dynamic load characteristics corresponding to the energy consumption source nodes and matched with the pre-configured equipment adjustment capabilities to calculate the transformation adaptability. Based on the transformation adaptability, target transformation equipment is selected and determined, including: The potential values ​​of the transformation are arranged into potential vectors according to the nodes in the energy consumption impact transmission chain. A back propagation matrix is ​​constructed based on the equipment topology. The importance weight of the nodes corresponding to the potential vector is determined based on the back attention mechanism. The weighted back tracing vector is obtained by combining the back propagation matrix. The energy contribution gradient is extracted from the weighted back tracing vector and the node corresponding to the gradient vanishing position is identified as the energy consumption source node. The load time series data of the equipment corresponding to the energy source node is extracted from the dynamic load characteristics. Wavelet packet decomposition is performed on the load time series data to obtain fluctuation components of different frequency bands. The dominant frequency band is identified from the fluctuation components and instantaneous energy is extracted. An energy envelope curve is constructed based on the instantaneous energy and the peak distribution pattern is extracted to obtain the fluctuation characteristics. The system acquires the pre-configured equipment adjustment capability and extracts the adjustment response bandwidth and energy buffer capacity. It extracts the frequency domain response requirement and instantaneous power surge from the fluctuation characteristics. It performs matching calculations between the frequency domain response requirement and the adjustment response bandwidth to obtain the frequency adaptation coefficient. It performs matching calculations between the instantaneous power surge and the energy buffer capacity to obtain the capacity adaptation coefficient. It then combines the frequency adaptation coefficient to solve for the modification adaptation degree. Based on the modification adaptation degree, it screens and determines the target modification equipment.

5. The method according to claim 4, characterized in that, The transformation potential values ​​are arranged into potential vectors according to the nodes in the energy consumption impact transmission chain. A backpropagation matrix is ​​constructed based on the equipment topology. The node importance weights corresponding to the potential vectors are determined based on the back attention mechanism, and the weighted back tracing vector is obtained by combining the backpropagation matrix. Extract the transformation potential value of each node from the energy consumption impact transmission chain and arrange them into a potential vector according to the node order. Extract the upstream dependency relationship between each energy-consuming device from the device topology relationship. Reverse the energy transmission direction in the upstream dependency relationship and construct a reverse connection graph. Calculate the node reverse propagation coefficient based on the in-degree and out-degree of each node in the reverse connection graph and fill it into a reverse propagation matrix according to the node position. The potential vector is used as the query vector, and the cumulative efficiency propagation value of each node is extracted from the energy consumption influence transmission chain as the key vector. The query vector and the key vector are multiplied by a dot product and a scaling transformation is applied to obtain the attention score. The attention score is normalized to obtain the attention weight distribution. The potential vector is weighted and summed based on the attention weight distribution to obtain the context-aware potential vector. The context-aware potential vector and the potential vector are residually connected and adjusted by a gating unit to obtain the node importance weight. The node importance weights are converted into diagonal matrices and multiplied with the backpropagation matrix to obtain a weighted backpropagation matrix. The potential vector is then multiplied with the weighted backpropagation matrix using iterative matrix multiplication, and the results are accumulated to obtain a weighted backtracking vector.

6. The method according to claim 1, characterized in that, Based on the dynamic load characteristics and preset process constraints, an operating state constraint space corresponding to the target modified equipment is constructed, and the energy consumption parameters are optimized and solved to obtain energy-saving operating parameters, including: Based on the dynamic load characteristics, the load change rate is determined. The process flow stability boundary and capacity guarantee threshold are extracted from the preset process constraints. The dynamic constraint boundary is determined based on the load change rate and the process flow stability boundary, and the static constraint boundary is determined based on the capacity guarantee threshold. The intersection operation of the dynamic constraint boundary and the static constraint boundary is performed to obtain the feasible operating domain. Based on the feasible operating domain, a set of candidate operating points is divided and the energy input to output efficiency ratio of each candidate operating point is calculated to obtain an initial efficiency score. Downstream related devices are extracted from the energy consumption influence transmission chain and the energy consumption transmission influence of each candidate operating point on the downstream related devices is calculated. The energy consumption transmission influence and the initial efficiency score are weighted and summed to obtain a global efficiency score. Based on the global efficiency score, a subset of high-efficiency operating points is obtained. The operating points in the subset of high-efficiency operating points are time-series aligned with the dynamic load characteristics, and the time offset is calculated to obtain the response delay coefficient. The global efficiency score and the response delay coefficient are normalized and weighted to obtain a comprehensive evaluation index. Based on the comprehensive evaluation index, the optimal operating point is selected and the corresponding control parameter combination is extracted to obtain energy-saving operating parameters.

7. The method according to claim 1, characterized in that, The energy-saving operating parameters are forward-propagated along the energy consumption impact transmission chain to calculate the parameter adjustment amount for each associated device. Based on the parameter adjustment amount and the device topology relationship, a collaborative transformation configuration is generated and a transformation plan is output, including: Extract the control parameter combination and energy input change value of the target modified equipment from the energy-saving operation parameters, extract the propagation direction and energy transfer coefficient in the energy consumption influence transmission chain, propagate the energy input change value to downstream related equipment layer by layer and accumulate to obtain the energy consumption change of each related equipment, and perform reverse calculation based on the energy consumption change and the equipment efficiency evaluation value of each related equipment to obtain the parameter adjustment amount of each related equipment. The parallel and series relationships between devices are extracted from the device topology. Based on the parallel relationship, the parameter adjustment amount in the parallel device group is load balanced and distributed to obtain the balanced adjustment amount. Based on the series relationship, the parameter adjustment amount in the series device chain is time-series coordinated and corrected to obtain the coordinated adjustment amount. The control parameters of the target equipment to be modified, along with the equalization adjustment amount and the coordination adjustment amount, are combined according to the equipment hierarchy in the equipment topology to construct a collaborative modification configuration and output a modification scheme.

8. An intelligent generation system for industrial energy-saving renovation schemes, used to implement the method of any one of claims 1-7, characterized in that, include: The coupling calculation unit is used to obtain the operating parameters and energy consumption parameters of each energy-consuming device in the industrial system, perform time-domain decomposition on the operating parameters to obtain dynamic load characteristics, construct the device topology relationship based on the energy consumption parameters, and perform correlation calculation with the dynamic load characteristics to obtain the device efficiency evaluation value and coupling dependency. The propagation analysis unit is used to identify key energy consumption paths based on the coupling dependency and perform propagation analysis on the equipment efficiency evaluation value to obtain the energy consumption impact transmission chain. Based on the equipment efficiency evaluation value of each node in the energy consumption impact transmission chain, the unit performs adjustability determination to obtain the transformation potential value. The source matching unit is used to identify energy consumption source nodes by tracing back along the energy consumption impact transmission chain based on the transformation potential value, determine the fluctuation characteristics based on the dynamic load characteristics corresponding to the energy consumption source nodes and match them with the pre-configured equipment adjustment capabilities to calculate the transformation adaptability, and screen and determine the target transformation equipment based on the transformation adaptability. The scheme generation unit is used to construct the operating state constraint space corresponding to the target modification equipment based on the dynamic load characteristics and preset process constraints, optimize and solve the energy consumption parameters to obtain energy-saving operating parameters, forward propagate the energy-saving operating parameters along the energy consumption influence transmission chain to calculate the parameter adjustment amount of each associated equipment, generate a collaborative modification configuration based on the parameter adjustment amount and the equipment topology relationship, and output the modification scheme.

9. An electronic device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to invoke instructions stored in the memory to execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, they implement the method described in any one of claims 1 to 7.

Images