Data-driven industrial circulating cooling water waste heat multi-source coupling system

Abstract

The present application relates to the technical field of industrial waste heat recovery optimization, in particular to an industrial circulating cooling water waste heat multi-source coupling system based on data driving, which comprises a waste heat collection unit, a data processing unit and an intelligent coupling regulation unit; in the present application, the temperature, flow and pressure data are obtained by the sensing assembly of the waste heat collection unit; after the data processing unit pre-processes the data, the temperature is divided into high, medium and low temperature layers, the heat value sublayer is subdivided by combining the clustering algorithm, the dynamic characteristic archives containing the parameter change curve, fluctuation amplitude and duration are established, the archives are compared with the pre-stored user heat demand characteristics by the intelligent coupling regulation unit, the matching degree is calculated according to the temperature coincidence degree, the single or multi-source collaborative coupling path is generated, the target heat parameter is synthesized by the heat capacity weighting in the multi-source mode, the waste heat supply and demand matching accuracy and utilization efficiency are improved, the diversified energy demand of industrial scene is adapted, and the energy saving and emission reduction are assisted.

Description

Technical Field

[0001] This invention relates to the field of industrial waste heat recovery optimization technology, and more specifically, to a data-driven multi-source coupling system for industrial circulating cooling water waste heat. Background Technology

[0002] Industrial waste heat recovery optimization is an important technology. The large amount of waste heat contained in industrial circulating cooling water has become an important energy-saving resource. Through efficient recovery and cascade utilization, it is of key significance for reducing enterprise energy consumption, improving energy utilization efficiency, and reducing carbon emissions. This technology breaks through the limitations of the traditional single waste heat utilization mode and provides intelligent solutions for matching waste heat supply and demand in multiple scenarios. The traditional waste heat distribution method that relies on manual experience is no longer able to meet the refined energy needs of complex industrial scenarios.

[0003] However, existing industrial circulating cooling water waste heat utilization technologies suffer from a core problem: insufficient precision in matching waste heat supply and demand. Traditional solutions do not classify waste heat according to parameters such as temperature and flow rate, nor do they analyze its dynamic characteristics. Instead, they simply distribute waste heat through fixed paths. When user heat demand does not match the characteristics of a single waste heat source, it can easily lead to an oversupply or undersupply of heat. At the same time, the lack of a multi-source waste heat synergistic coupling mechanism makes it impossible to compensate for the deficiencies of a single heat source through weighted allocation of waste heat at different levels. This results in waste heat resource waste and a decline in user energy experience. These combined problems lead to low waste heat utilization efficiency, making it difficult to adapt to the diverse and dynamic energy demands of industrial scenarios and hindering the achievement of energy conservation and emission reduction goals. To solve this technical problem, we provide a data-driven multi-source coupling system for industrial circulating cooling water waste heat. Summary of the Invention

[0004] The purpose of this invention is to provide a data-driven multi-source coupling system for waste heat from industrial circulating cooling water, in order to solve the problems mentioned in the background art.

[0005] 1. Because traditional solutions do not classify and dynamically analyze waste heat and use fixed path allocation, supply and demand mismatch occurs. Therefore, this case uses a data processing unit to divide waste heat into levels and establish dynamic characteristic profiles, and an intelligent coupling control unit to generate matching coupling paths, which can improve the accuracy of waste heat supply and demand matching.

[0006] 2. Since traditional solutions lack a multi-source coordination mechanism and cannot compensate for the defects of a single heat source, resulting in resource waste, this case uses an intelligent coupling control unit to initiate a multi-source coordination mode, weighted synthesis of target thermal parameters to generate a coordination path, which can improve waste heat utilization and meet diverse needs.

[0007] To achieve the above objectives, a data-driven multi-source coupling system for waste heat from industrial circulating cooling water is provided, including:

[0008] The waste heat acquisition unit transmits monitoring data to the data processing unit through sensing components that monitor temperature, flow rate, and pressure parameters;

[0009] The data processing unit preprocesses the monitoring data, then divides the preprocessed monitoring data into three levels, and establishes dynamic characteristic files for waste heat at each level. The dynamic characteristic files include parameter change curves, parameter fluctuation amplitudes, and durations.

[0010] The intelligent coupling control unit has a built-in self-learning database and coupling path optimization module. The self-learning database pre-stores the heat demand characteristic parameters of waste heat users in different industrial scenarios, including the demand temperature range, heat load stability and continuous heat use duration.

[0011] The coupling path optimization module compares the dynamic characteristic profile with the heat demand characteristic parameters and calculates the matching degree between the two. When calculating the matching degree, the overlap of temperature range is compared first, and then the adaptability of flow fluctuation and the consistency of duration are compared in turn. When the matching degree is higher than the preset threshold, the coupling path between the corresponding level of waste heat and user demand is generated.

[0012] When the matching degree is lower than the preset threshold, the coupling path optimization module starts the multi-source collaborative coupling mode, that is, selects two waste heat sources at different levels, calculates the adjustable heat capacity according to the real-time parameters of each waste heat source, and synthesizes the target heat parameters according to the heat capacity ratio, so that the matching degree between the target heat parameters and the user demand characteristic parameters reaches above the preset threshold, and generates a collaborative coupling path that includes the output ratio of each waste heat source and the transmission path switching node.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0014] 1. The data processing unit divides waste heat into high, medium, and low temperature layers based on temperature ranges. It combines clustering algorithms to perform secondary clustering of flow and pressure parameters to form calorific value sub-layers. Then, it extracts dynamic features through time series neural networks to establish dynamic characteristic profiles that include parameter change curves, fluctuation amplitudes, and durations. This enables a refined characterization of waste heat resources. The intelligent coupling control unit compares these profiles with user heat demand features pre-stored in the self-learning database. It calculates the matching degree according to the priority of temperature overlap, flow adaptability, and duration fit, and generates targeted coupling paths. This avoids the oversupply or undersupply of heat caused by traditional fixed allocation and significantly improves the accuracy of supply and demand matching.

[0015] 2. When the matching degree of a single waste heat source is insufficient, the system excludes the levels with excessive temperature deviation, selects combinations that meet the required redundancy in heat capacity and whose high-temperature levels reach the minimum guaranteed value, calculates the adjustable heat capacity based on real-time parameters, and synthesizes the target heat parameters by weighting according to proportion. The temperature adopts linear fusion and the flow rate is smoothed to make the fluctuation of the synthesized parameters lower than that of a single source. At the same time, a collaborative path including output ratio and path switching nodes is generated, and stable transmission is achieved through intelligent regulating valve group and pressure buffer device. This not only makes up for the defects of a single heat source, but also fully explores the utilization value of waste heat at different levels and reduces resource waste.

[0016] 3. The digital twin platform simulates and verifies the coupling path, pre-adjusts valve opening to ensure feasibility, and uses a pipeline attenuation compensation mechanism to correct heat capacity loss during long-distance transmission, thereby improving data accuracy. The self-learning database continuously accumulates demand characteristics and matching cases from different scenarios, supporting the system to dynamically optimize matching logic and enhancing its adaptability to diverse and dynamic chemical conditions. In addition, multi-level response commands in the collaborative path enable time-sequential switching, avoiding fluid impact, ensuring transmission stability, and ultimately improving overall waste heat utilization efficiency, thus helping enterprises achieve their energy conservation and emission reduction goals. Attached Figure Description

[0017] Figure 1 This is an overall block diagram of the present invention.

[0018] The meanings of the labels in the diagram are as follows:

[0019] 1. Waste heat collection unit; 2. Data processing unit; 3. Intelligent coupling and control unit. Detailed Implementation

[0020] 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.

[0021] This invention provides a data-driven multi-source coupling system for waste heat from industrial circulating cooling water. Please refer to [link / reference]. Figure 1 As shown, it includes:

[0022] Waste heat acquisition unit 1 transmits monitoring data to data processing unit 2 through sensing components that monitor temperature, flow rate, and pressure parameters;

[0023] Data processing unit 2 preprocesses the monitoring data, then divides the preprocessed monitoring data into three levels, and establishes dynamic characteristic files for waste heat at each level. The dynamic characteristic files include parameter change curves, parameter fluctuation amplitudes, and durations.

[0024] In industrial circulating cooling water systems, waste heat acquisition unit 1 is the source of data acquisition. It collects monitoring data reflecting waste heat characteristics in real time through sensing components deployed at key nodes of the circulating cooling water network. These data specifically include the cooling water temperature, real-time flow rate, and pressure value in the pipeline at each monitoring point. These parameters are directly related to the energy level and transmission status of waste heat and are the basis for subsequent data processing and coupled control.

[0025] The collected monitoring data is first transmitted to data processing unit 2 for preprocessing to ensure data integrity and consistency, providing reliable input for subsequent hierarchical division. After preprocessing, data processing unit 2 begins to perform hierarchical division of the monitoring data. The specific method for dividing the preprocessed monitoring data into three levels is as follows:

[0026] This process begins with temperature, as temperature is a core indicator for measuring the quality of waste heat energy. The waste heat source here comes from the circulating cooling water discharged from various equipment during industrial production. This cooling water carries varying degrees of heat after exchanging heat with the equipment, forming a recyclable waste heat source. Based on a preset temperature range, the waste heat source is divided into high-temperature, medium-temperature, and low-temperature layers. The high-temperature layer (60-90℃) is used for scenarios requiring high-temperature heat sources, such as heating and raw material preheating. The medium-temperature layer (30-60℃) is used for process insulation and hot water supply. The low-temperature layer (20-30℃) is used for low-grade heat applications such as agricultural greenhouses and fishpond insulation. This division method clarifies the potential application directions based on the energy quality of the waste heat, laying the foundation for precise matching of user needs. After completing the temperature layer division, to further refine the differences in waste heat within the same temperature layer, a secondary clustering algorithm is used to cluster the flow and pressure parameters of the same temperature layer. A dynamic density peak clustering algorithm is employed, and the specific process is as follows:

[0027] The flow rate and pressure parameters of all monitoring data within the same temperature level are used as two-dimensional feature vectors. Each data point can be represented as (flow rate, pressure). The local density of each data point is calculated, which is the number of data points within a certain distance around the point and the minimum distance between the data point and the data point with higher density. Based on this, the cluster center is determined. Points with high local density and far distance from other high-density points are selected as centers. Data points are classified according to their distance from each cluster center. The data point closest to a certain center is assigned to that cluster, forming multiple subgroups with similar flow rate-pressure characteristics. For example, in the mesothermal layer (30-60℃), there exists a cluster with stable flow rate (20-30 m³ / h) and small pressure fluctuation (0.3-0.4 MPa), and another cluster with large flow rate fluctuation (10-40 m³ / h) and higher pressure (0.5-0.6 MPa). Secondary clustering can capture the differences in waste heat transfer characteristics within the same temperature range, making the hierarchical division more refined. Then, based on the distance between the data points—that is, each monitoring data sample containing flow and pressure parameters—and the cluster center, different calorific value sub-layers are defined. Specifically, the ohmic distance from each data point to its respective cluster center is calculated. The distance is calculated by dividing data points into core sub-layers based on the average distance from all data points within a cluster to the center. Data points with a distance less than or equal to a preset threshold (which is the average distance from all data points to the center of the cluster) represent the most stable waste heat portion within that cluster. Data points with a distance greater than the threshold but less than twice the threshold are divided into edge sub-layers. Through this division, each temperature level is further subdivided into multiple calorific value sub-layers, ultimately forming a waste heat hierarchy system that includes "temperature level + calorific value sub-layers". Each sub-layer is labeled with a dynamic tag containing key information such as the flow and pressure fluctuation range and the average calorific value of that sub-layer, facilitating rapid identification and matching by the subsequent intelligent coupling control unit 3.

[0028] Methods for establishing dynamic characteristic profiles of waste heat at each level include:

[0029] After completing the waste heat stratification and labeling with dynamic tags, a dynamic characteristic profile needs to be established for each stratum to fully understand its variation patterns. This process achieves in-depth characterization of waste heat characteristics through time series analysis and pattern comparison. A time series neural network is used to extract features from the monitoring data using an adaptive window. The adaptive window automatically adjusts its size based on data fluctuations. Extracted features include: the mean value of the parameter within the window (reflecting the average level during that period), peak and trough values ​​(reflecting the amplitude of parameter fluctuations), trend slope (indicating the direction of parameter change), and fluctuation frequency (statistically, the number of times the parameter fluctuates per unit time). These features capture the dynamic variation characteristics of waste heat parameters from different dimensions, providing structured data for subsequent analysis. After feature extraction, due to these... Features reflect the historical variation patterns of parameters, and prediction models need to infer future trends based on historical patterns. Therefore, the extracted features are input into the prediction model (a sub-module of a time-series neural network, using an LSTM structure to capture temporal dependencies). By learning information such as trend slope and fluctuation frequency in the features, the model predicts the fluctuation trend of parameters within the next 30 minutes, enabling it to anticipate the direction of change in waste heat parameters in advance and provide a forward-looking basis for coping with fluctuations. Real-time fluctuation data is dynamically compared with historical patterns, and abnormal conditions are marked. Real-time fluctuation data refers to the raw parameter values ​​collected within the current monitoring period and the features calculated therefrom. Historical patterns are typical parameter change patterns stored in the historical data accumulated by the system, categorized by waste heat level and sub-level. The specific comparison process is as follows:

[0030] The similarity between the features of real-time fluctuation data and the features of corresponding historical patterns is calculated, and a similarity threshold is set. When the similarity is lower than the threshold, the source of deviation is further analyzed. If the deviation exceeds the temperature range of that level, it is marked as an abnormal condition, and information such as the time of occurrence of the abnormality and the parameter deviation value is recorded. At the same time, the probability characteristic envelope of the parameter change curve is generated. The innovation lies in combining the predicted trend of the prediction model with the historical fluctuation probability distribution. The specific process is as follows:

[0031] Based on the real-time parameter change curve, and the distribution of fluctuations in the waste heat parameters at this level in historical data, probability boundaries are drawn on both sides of the curve. The upper boundary is the predicted trend value plus the upper limit of the historical 95% confidence interval, and the lower boundary is the predicted trend value minus the lower limit of the historical 95% confidence interval, forming an envelope that encloses the parameter change curve. This envelope can intuitively reflect the possible fluctuation range of the parameters in the future, providing a visual basis for judging whether the parameters are within the normal fluctuation range, making subsequent supply and demand matching more targeted, and effectively improving the system's adaptability to dynamic changes in waste heat.

[0032] The intelligent coupling control unit 3 has a built-in self-learning database and coupling path optimization module. The self-learning database pre-stores the heat demand characteristic parameters of waste heat users in different industrial scenarios, including the demand temperature range, heat load stability and continuous heat use duration.

[0033] After establishing dynamic characteristic profiles for waste heat at each level, to accurately measure the degree of matching between waste heat and user needs, the matching degree needs to be calculated from three dimensions: temperature, flow rate, and duration. Among these, temperature range overlap, flow rate fluctuation adaptability, and duration fit are the core evaluation indicators. First, let's look at temperature range overlap, which is defined as the overlap ratio between the user's required temperature range and the waste heat source temperature fluctuation range. The user's required temperature range is the temperature range required by the user in a certain industrial scenario, pre-stored in the self-learning database. The waste heat source temperature fluctuation range comes from the temperature parameter change curve of that waste heat level in the dynamic characteristic profile, that is, the interval formed by the maximum and minimum temperatures over a period of time. The overlap ratio is calculated as follows:

[0034] First, determine the overlapping portion of the two intervals (in the example above, the overlapping interval is 40-50℃). Then, divide the length of the overlapping interval (10℃) by the length of the user's required temperature range (10℃) to obtain a 100% overlap. If the user's required temperature range is 50-60℃, and the overlap with the waste heat source interval of 5-55℃ is 50-55℃ (length 5℃), then the overlap is 0%. This indicator directly reflects the basic compatibility between waste heat temperature and user needs, and is the primary consideration in the matching degree calculation. The flow fluctuation adaptability is calculated using a dynamic morphological matching algorithm to determine the similarity between the waste heat source flow curve and the user's demand baseline. The waste heat source flow curve is extracted from the dynamic characteristic file and is a continuous curve drawn from the flow monitoring data of a certain waste heat level within a unit of time (the flow value is recorded every 5 minutes, and connected to form a 24-hour flow change curve). The user demand baseline is a flow reference curve generated based on the user's heat load stability requirements (if a user needs a stable flow, the baseline is a horizontal straight line; if the user's demand changes periodically over time, the baseline is the corresponding periodic curve). The specific calculation process is as follows:

[0035] The two curves are aligned along the time axis, and a dynamic time warping algorithm (introducing a time elastic matching mechanism, allowing the curves to locally expand and contract in the time dimension) is used to calculate the distance between them. The smaller the distance, the higher the similarity. For example, the waste heat source flow rate curve shows a small fluctuation (±2m³ / h) in a certain period of time, while the user demand baseline is a horizontal straight line (target flow rate 20m³ / h). The algorithm calculates the sum of squared deviations at each point by aligning the fluctuation nodes with the baseline value and takes the average to obtain a similarity score (out of 100 points; if the average deviation is 1m³ / h, the score may be 90 points). This score is the quantitative result of the flow rate fluctuation adaptability, which can reflect whether the waste heat flow rate can adapt to the user's load change demand. The duration of fit is the degree to which the sustainable heating duration of the waste heat source covers the duration of user demand. The sustainable heating duration of the waste heat source comes from the parameter duration record in the dynamic characteristic file, that is, the continuous duration for which the waste heat level can stably output heat that meets the basic temperature and flow requirements. The user demand duration is the time when the user needs to use heat continuously in the self-learning database. The calculation of the coverage is as follows:

[0036] If the sustainable duration of the waste heat source (8 hours) is greater than or equal to the user's demand duration (6 hours), the matching degree is 100%. If the sustainable duration of the waste heat source is 4 hours and the user's demand duration is 6 hours, the matching degree is 4 / 6≈67%. This indicator ensures that the waste heat supply can meet the user's continuous energy demand and avoid production disruptions due to heating interruptions. It provides a scientific basis for the intelligent coupling control unit 3 to generate accurate coupling paths, making the waste heat supply and demand matching more in line with the actual needs of industrial scenarios.

[0037] The coupling path optimization module compares the dynamic characteristic profile with the heat demand characteristic parameters and calculates the matching degree between the two. When calculating the matching degree, the overlap of temperature range is compared first, and then the adaptability of flow fluctuation and the consistency of duration are compared in turn. When the matching degree is higher than the preset threshold, the coupling path between the corresponding level of waste heat and user demand is generated.

[0038] The matching degree is calculated as follows:

[0039] After completing individual calculations for temperature range overlap, flow fluctuation adaptability, and duration fit, the coupled path optimization module performs quantitative analysis of the system by constructing a multi-dimensional feature vector space to comprehensively evaluate the overall matching level between waste heat and user needs.

[0040] A multi-dimensional feature vector space is constructed, specifically using temperature range overlap, flow rate fluctuation adaptability, and duration fit as three dimensions. The numerical range of each dimension is standardized to 0-1 (1 represents a perfect match). The corresponding parameters in the dynamic characteristic profile and the user's heat demand feature parameters are respectively transformed into feature vectors in this space. For example, the vector of a waste heat source is (0.8, 0.7, 0.9), representing 80% temperature overlap, 70% flow rate adaptability, and 90% duration fit, while the user demand vector is (1.0, 1.0, 1.0) (ideal state). The role of the multidimensional feature vector space is to transform abstract matching indicators into quantifiable spatial coordinates, allowing the differences in matching between different waste heat sources and user needs to be intuitively reflected through vector distance. This provides a geometric basis for subsequent similarity calculations. A similarity algorithm is used to calculate the matching value and map it to a standard interval. Specifically, the cosine similarity algorithm is used to calculate the cosine value of the angle between two feature vectors (ranging from -1 to 1). The closer the cosine value is to 1, the more consistent the directions of the two vectors are, and the higher the matching degree. For example, the cosine similarity between the waste heat source vector and the user need vector is (0.8 × 1.0). + 0.7×1.0+0.9×1.0)÷(√(0.8²+0.7²+0.9²)×√(1.0²+1.0²+1.0²))≈0.95, and then the cosine value is converted into a standard interval score of 0-100 through linear mapping (e.g., 0.95 corresponds to 95 points). The mapping formula is "standard score = (cosine value + 1) ÷ 2×100", ensuring that the matching value can reflect the relative difference and is easy to understand intuitively. This process condenses the matching degree of multi-dimensional features into a single quantitative index, providing a basis for quickly screening qualified waste heat sources. When the matching values ​​of multiple waste heat sources are all higher than the preset threshold, in order to further screen the optimal solution, a multi-objective optimization screening mechanism is introduced to perform dual priority ranking and construct an "efficiency-stability" dual-objective optimization model:

[0041] The first ranking prioritizes waste heat utilization efficiency (calculating the proportion of waste heat converted into effective work), sorting by efficiency from high to low. The second ranking prioritizes system operational stability (based on a comprehensive evaluation of flow fluctuation adaptability and historical fault frequency), sorting waste heat sources with similar efficiency by stability from high to low. For example, if waste heat sources A and B both have a matching score of 90, with A having a utilization efficiency of 85% and a stability score of 90, and B having a utilization efficiency of 82% and a stability score of 95, then A takes precedence over B in the first ranking. If prioritizing stability is required, B can be moved forward in the comprehensive ranking through weight adjustments. After ranking, a list of non-dominated solutions is output, eliminating obviously inferior solutions (such as those with lower efficiency and stability than other solutions), retaining all solutions that have advantages in at least one objective, and labeling the efficiency value, stability score, and applicable scenarios of each solution (e.g., A is suitable for high-efficiency priority scenarios, B is suitable for stability priority scenarios). This provides flexible space for subsequent coupling path selection, lays a scientific foundation for the intelligent coupling control unit 3 to generate the optimal coupling path, and effectively improves the accuracy and flexibility of waste heat supply and demand matching.

[0042] The logic for generating the coupling path is as follows:

[0043] After determining the matching degree between waste heat and user demand and selecting the optimal waste heat source, the intelligent coupling control unit 3 needs to generate specific coupling paths to achieve efficient transmission of waste heat from the source to the user. The industrial scenario topology map is a digital representation of the spatial distribution and connection relationships of waste heat supply points, transmission pipelines, user energy consumption points, and related equipment within the industrial plant area. It marks the location of each waste heat source, the distribution of user demand points, the direction and specifications of pipelines, the control range of valves, and other information, which is equivalent to a "digital map" of the entire industrial waste heat transmission system, providing a spatial basis for path planning. Based on this topology map, the intelligent coupling control unit 3 first selects the waste heat source with the highest matching degree as the main supply source, and then determines the transmission path with the lowest energy consumption through a path search algorithm, which uses dynamic energy consumption weights. algorithm:

[0044] The algorithm starts with the waste heat source and ends with the user's demand point. It decomposes the energy consumption of the transmission path into pipe friction heat loss (related to pipe length and insulation coefficient) and pump energy consumption (related to flow rate and pipe resistance). Each pipe segment is assigned a dynamic weight (a weighted sum of heat loss and transmission energy consumption). During the search process, it not only prioritizes pipe segments with low weight values ​​but also avoids high-resistance segments (such as abrupt changes in pipe diameter) and high-loss segments (such as areas with damaged insulation) marked on the topology map in real time. By continuously iterating and updating the path nodes, it ultimately generates the transmission path with the lowest total energy consumption. For example, there are two paths, A and B, from the high-temperature waste heat source to a user. Path A is shorter but has poor pipe insulation (large heat loss), while path B is slightly longer but has good insulation and low pump energy consumption. The algorithm calculates that path B has lower total energy consumption and selects it as the optimal path. After determining the transmission path, the opening of the intelligent regulating valve group needs to be pre-adjusted to match the user's real-time heat demand, because the valve opening in the path directly affects the flow rate and pressure, thus affecting the stability of the heating supply. The specific process is as follows:

[0045] Based on the user's required flow rate and the total path resistance, the target opening degree required for each valve is calculated. Initially, a 50% opening degree is pre-adjusted. Then, the current flow rate is fed back in real time by flow sensors installed after the valves. The opening degree is fine-tuned after comparing this with the target flow rate. This adjustment is repeated until the flow rate stabilizes within ±5% of the target value. This pre-adjustment process ensures stable parameters at the start of the path, avoiding thermal shock caused by sudden flow changes. To further verify the feasibility of the path, simulation verification is performed using a digital twin platform. This platform constructs a virtual model completely consistent with the physical system, including the thermal conduction characteristics of the pipeline, the adjustment response of the valves, and the dynamic changes in user demand. During simulation, the pre-adjusted valve opening degree and the thermal parameters of the main power source are input. The virtual model simulates the transmission process over the next hour, monitoring for issues such as overpressure, overheating loss, and unstable flow. If problems are found, feedback is sent to the intelligent coupling control unit 3 to re-optimize the path or adjust the valve opening until the simulation results show that all parameters are within a safe range. For example, if the simulation finds that the original path experiences flow fluctuations exceeding the user's tolerance range at the 30-minute mark, the platform will automatically suggest adjusting the opening curve of a certain valve. After a second simulation verification, a feasible coupling path is finally determined, ensuring that the generated coupling path accurately adapts to user needs and the actual conditions of the industrial scenario. This provides a reliable guarantee for the efficient transmission of waste heat and lays the foundation for the single-source path in the subsequent multi-source collaborative coupling mode.

[0046] When the matching degree is lower than the preset threshold, the coupling path optimization module starts the multi-source collaborative coupling mode, that is, selects two waste heat sources at different levels, calculates the adjustable heat capacity according to the real-time parameters of each waste heat source, and synthesizes the target heat parameters according to the heat capacity ratio, so that the matching degree between the target heat parameters and the user demand characteristic parameters reaches above the preset threshold, and generates a collaborative coupling path that includes the output ratio of each waste heat source and the transmission path switching node.

[0047] The strategy for selecting two waste heat sources at different levels is as follows:

[0048] When the matching degree of a single waste heat source is lower than a preset threshold, the coupling path optimization module of the intelligent coupling control unit 3 will activate the multi-source collaborative coupling mode. First, two waste heat sources at different levels need to be selected. This process follows a strategy that combines precise screening with system evaluation, excluding waste heat sources whose temperature range exceeds the preset deviation. The target temperature range is the core temperature range required by the user. The preset deviation is usually set to ±10℃. If the temperature fluctuation range of a certain high-temperature level is 70-80℃, and the deviation from the target range reaches 20℃, it will be excluded to ensure that the selected waste heat sources have basic compatibility with the user's needs in terms of temperature attributes, laying the foundation for subsequent thermal parameter synthesis. In the remaining levels that meet the temperature deviation requirements, a heat capacity contribution evaluation matrix is ​​constructed. The specific process is as follows:

[0049] An initial matrix is ​​formed using waste heat source levels (such as high-temperature sub-layers, mid-temperature sub-layers, etc.) as rows and heat capacity-related indicators as columns (including instantaneous heat capacity, heat capacity decay rate during continuous heating, and heat loss rate during transmission). Each indicator is then assigned a weight (instantaneous heat capacity weight 0.5, decay rate weight 0.3, heat loss rate weight 0.2), and the weighted score for each level is calculated as its heat capacity contribution. For example, a high-temperature sub-layer may have a high instantaneous heat capacity but decays quickly, while a mid-temperature sub-layer may have a slightly lower instantaneous heat capacity but decays slowly. The matrix quantifies the combined contribution of both. Its function is to transform the heat capacity characteristics of different levels into comparable quantitative data, avoiding the bias caused by relying solely on a single indicator to select waste heat sources. Based on the evaluation matrix, combinations of waste heat sources whose effective heat capacity meets the redundancy requirements of waste heat user demand and whose high-temperature level proportion reaches the minimum guaranteed value are selected. Effective heat capacity refers to the actual usable heat after deducting transmission losses. Demand redundancy is typically... The minimum guaranteed percentage of the high-temperature heat source is set at 10% (meaning the effective heat capacity must reach 110% of the user's required heat capacity to cope with fluctuations). The minimum guaranteed percentage of the high-temperature heat source is set at 30% (meaning the heat capacity provided by the high-temperature heat source in the combination must not be less than 30% of the total effective heat capacity to ensure the temperature basis of the combined heat source). For example, if the user's required heat capacity is 100kW, and a combination consists of a high-temperature heat source (effective heat capacity 40kW) and a medium-temperature heat source (effective heat capacity 70kW), the total effective heat capacity is 110kW (meeting 10% redundancy), and the high-temperature heat source accounts for 36.4% (exceeding the 30% guaranteed value). This combination meets the screening criteria. If another combination meets the standard for total effective heat capacity but the high-temperature heat source accounts for only 25%, it will be excluded. This allows the two selected heat sources to complement each other in synergistic coupling, providing a reliable source guarantee for subsequent weighted synthesis of target heat parameters and generation of efficient synergistic paths. It also enables the multi-source mode to truly make up for the defects of a single heat source and improve the flexibility and stability of waste heat utilization.

[0050] The method for calculating adjustable heat capacity is as follows:

[0051] After selecting two waste heat sources at different levels, to ensure the accuracy of heat capacity allocation during multi-source synergistic coupling, it is necessary to calculate the adjustable heat capacity of each waste heat source. This process relies on the dynamic updating of real-time data and also needs to consider heat loss during transmission. Based on the real-time collected temperature and flow data, the instantaneous heat capacity value of each waste heat source is dynamically refreshed. Specifically, the formula for calculating the instantaneous heat capacity value is "Heat capacity = Flow rate × Temperature × Specific heat capacity × Density" (where specific heat capacity and density are inherent physical parameters of circulating cooling water, preset as constants). Waste heat acquisition unit 1 collects the real-time temperature (e.g., the current temperature of the high-temperature layer waste heat source is 75℃) and flow rate (e.g., 30 m³ / h) of each waste heat source every 5 minutes. Substituting these data into the formula, the current heat capacity is calculated (e.g., 30 × 75 × 4.2 × 1000 ≈ 9.45 × The system automatically overwrites the values ​​from the previous cycle, achieving dynamic updates. This process reflects the instantaneous heating capacity of the waste heat source in real time, providing the latest basis for subsequent weighted synthesis. Since multi-source collaborative coupling may involve long-distance transmission, heat loss occurs during transmission due to heat dissipation from the pipelines. Therefore, a pipeline attenuation compensation mechanism needs to be introduced to pre-correct the heat capacity loss. The specific process is as follows:

[0052] First, determine the transmission pipe length from each waste heat source to the user's demand point based on the industrial scenario topology diagram (e.g., the transmission distance for the waste heat source in the high-temperature layer is 500 meters). Then, call the pipe attenuation coefficient table (this table presets attenuation rates for different distances based on pipe material, insulation layer thickness, and ambient temperature; for example, ordinary steel pipes attenuate by 2% per 100 meters) to calculate the total attenuation rate (10% attenuation for 500 meters). Multiply the instantaneous heat capacity value by (1 - total attenuation rate) to obtain the corrected adjustable heat capacity (e.g., 9.45 × 10 ... ×0.9≈8.505× The effect of the correction is as follows:

[0053] This makes the calculated adjustable heat capacity closer to the actual heat reaching the user end, avoiding insufficient heat supply due to not considering attenuation (for example, without correction, the actual heat reaching the user end may only be 90% of the demand if the original heat capacity is adjusted, but after correction, the attenuation part can be made up in advance to ensure supply and demand balance). This makes the calculation of adjustable heat capacity more in line with actual operating conditions, providing accurate basic data for subsequent weighted synthesis of target heat parameters based on heat capacity ratio, and also enabling the multi-source collaborative coupling mode to more accurately meet user needs, effectively improving the efficiency of waste heat transfer and utilization.

[0054] The logic for weighted synthesis of target thermal parameters is as follows:

[0055] After determining the adjustable heat capacity of each waste heat source, the process of weighted synthesis of target thermal parameters needs to consider the synergistic characteristics of temperature and flow rate to achieve precise matching with user needs. The heat capacity proportion of different levels of waste heat sources is calculated as a weighting coefficient, i.e., the heat capacity proportion of a certain waste heat source = the adjustable heat capacity of that waste heat source ÷ the sum of the adjustable heat capacities of all waste heat sources participating in the synergy (e.g., the adjustable heat capacity of the waste heat source in the high-temperature layer is 8.5 × 10⁻⁶). The mesothermal layer is 6.3× If the high-temperature layer accounts for approximately 57% and the medium-temperature layer accounts for 43%, this proportion directly reflects the energy contribution weight of each waste heat source in the synergistic coupling. For temperature parameters, a linear weighted fusion is used, specifically by multiplying the real-time temperature of each waste heat source by its corresponding heat capacity proportion and then summing the results to obtain the composite temperature (e.g., if the real-time temperature of the high-temperature layer is 75℃ and the medium-temperature layer is 50℃, then the composite temperature = 75 × 57% + 50 × 43% ≈ 64.7℃). This linear fusion method ensures that the composite temperature is within the user's required temperature range while retaining the temperature dominance of high-proportion waste heat sources and avoiding excessive impact of single heat source temperature fluctuations on the overall temperature. For flow parameters, a smoothing fusion is used, and a dynamic attenuation factor is introduced.

[0056] First, the initial weighted flow rate is calculated based on the heat capacity ratio (e.g., high-temperature layer flow rate 30 m³ / h, medium-temperature layer 25 m³ / h, initial weighted flow rate = 30 × 57% + 25 × 43% ≈ 27.85 m³ / h). Then, the real-time flow fluctuation of each waste heat source is monitored. If the flow fluctuation of a heat source exceeds the preset threshold, an attenuation factor is assigned to it (the larger the fluctuation, the smaller the factor, ranging from 0.8 to 1.0), and the weighted flow rate is recalculated (e.g., if the flow fluctuation of the medium-temperature layer reaches 6 m³ / h, the attenuation factor is 0.8, then the corrected weighted flow rate = 30 × 57% + 25 × 43% × 0.8 ≈ 26.1 m³ / h). Finally, the corrected flow rate for five consecutive cycles is smoothed using a moving average algorithm to obtain the final composite flow rate, ensuring that the stability of the output parameters meets user requirements and providing a reliable thermal parameter basis for the subsequent generation of collaborative coupling paths.

[0057] The process of generating a cooperative coupling path is as follows:

[0058] After weighted synthesis of the target thermal parameters, a specific collaborative coupling path needs to be generated to achieve efficient transmission and coordinated supply of multi-source waste heat. This process revolves around flow distribution, pressure buffering, and timing control to ensure the stability and accuracy of multi-source coupling. The flow distribution valve is driven according to the heat capacity ratio of each waste heat source. Specifically, an intelligent flow distribution valve is installed at the confluence pipe connecting two waste heat sources at different levels. After receiving the heat capacity ratio data (e.g., 57% for the high-temperature layer and 43% for the medium-temperature layer), the valve controller converts it into corresponding valve opening commands. The valve opening for the high-temperature layer is set at 57%, and the valve opening for the medium-temperature layer is set at 43%. At the same time, the actual flow ratio is monitored in real time by the flow sensor built into the valve. If the deviation from the target ratio exceeds 3% (e.g., the actual ratio of the high-temperature layer drops to 53%), the valve opening is automatically fine-tuned (increasing the valve opening of the high-temperature layer by 1%) until the flow distribution matches the heat capacity ratio. This consistent operation ensures that each waste heat source contributes to the coordinated supply as expected, providing a flow basis for the stable output of the target thermal parameters. Since fluid shocks may occur at the switching nodes of the transmission path, i.e., the confluence points of different heat source pipelines, due to pressure differences, a pressure buffer device needs to be deployed here. This device has a built-in elastic diaphragm and pressure sensor. When the pressure fluctuation in the pipeline exceeds ±0.1MPa, the diaphragm automatically expands and contracts to adjust its volume to absorb the pressure shock. If the pressure in the high-temperature layer rises sharply, the diaphragm expands to accommodate some fluid, reducing the pressure at the confluence point. Simultaneously, the pressure data is fed back to the control system. If the fluctuation lasts for more than 5 seconds, an auxiliary pressure relief valve is triggered for fine-tuning to further stabilize the pressure, preventing pipeline vibration or equipment damage due to shocks and ensuring the safety of the transmission process. Based on this, a control sequence containing multi-level response commands is generated to achieve time-sequential switching and parameter adjustment of the coupling path, specifically divided into three response levels:

[0059] The first-level command is the initial start-up phase (first 5 minutes), which controls the flow distribution valve to slowly open to an initial opening of 30%, while the pressure buffer device enters a pre-pressurization state to ensure smooth fluid convergence. The second-level command is the stable operation phase, which generates a fine-tuning command every 2 minutes based on the deviation between the real-time monitored target thermal parameters (temperature, flow rate) and user requirements (e.g., reducing the flow rate of the high-temperature layer by 2% when the temperature is too high). The third-level command is the switching transition phase, where, when a waste heat source needs to be removed or added, the pressure is first pre-adjusted through the buffer device, and then the flow distribution is adjusted according to the principle of "increase first, then decrease" (e.g., gradually increasing the proportion of the medium-temperature layer to 60% while simultaneously reducing the proportion of the high-temperature layer to 40%). The entire process uses timestamps to mark the execution order of each command, forming a time-sequential control sequence, which is stored in the command library of the intelligent coupling control unit 3 for the execution system to call in sequence, so that the synthesized target thermal parameters continuously meet user requirements, ultimately achieving multi-source collaborative and efficient utilization of waste heat from industrial circulating cooling water.

[0060] In this invention, the waste heat acquisition unit 1 acquires temperature, flow rate, and pressure data through sensing components. After the data processing unit 2 preprocesses the data, it divides the data into high, medium, and low temperature layers according to temperature, and further subdivides the calorific value sublayers using a clustering algorithm to establish a dynamic characteristic profile containing parameter change curves, fluctuation amplitudes, and durations. The intelligent coupling and control unit 3 compares the profile with pre-stored user heat demand characteristics, prioritizes calculating the matching degree based on temperature overlap, and generates single or multi-source collaborative coupling paths. In the multi-source mode, the target heat parameters are synthesized by weighted heat capacity to improve the accuracy and efficiency of waste heat supply and demand matching, adapt to the diverse energy needs of industrial scenarios, and contribute to energy conservation and emission reduction.

[0061] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A data-driven multi-source coupling system for waste heat from industrial circulating cooling water, characterized in that, include: The waste heat collection unit (1) transmits monitoring data to the data processing unit (2) through the sensing components that monitor temperature, flow rate and pressure parameters; The data processing unit (2) preprocesses the monitoring data, then divides the preprocessed monitoring data into three levels, and establishes dynamic characteristic files for waste heat at each level. The dynamic characteristic files include parameter change curves, parameter fluctuation amplitudes, and durations. The specific method for dividing the preprocessed monitoring data into three levels is as follows: Based on the preset temperature range, the waste heat source is divided into a high temperature layer, a medium temperature layer and a low temperature layer. The flow and pressure parameters of the same temperature level are clustered again by a clustering algorithm. Different calorific value sub-layers are divided according to the distance relationship between the data points and the cluster center to form waste heat levels and label them with dynamic tags. The intelligent coupling control unit (3) has a built-in self-learning database and coupling path optimization module. The self-learning database pre-stores the heat demand characteristic parameters of waste heat users in different industrial scenarios, including the demand temperature range, heat load stability and continuous heat use duration. The coupling path optimization module compares the dynamic characteristic profile with the heat demand characteristic parameters and calculates the matching degree between the two. When calculating the matching degree, the overlap of temperature range is compared first, and then the adaptability of flow fluctuation and the consistency of duration are compared in turn. When the matching degree is higher than the preset threshold, the coupling path between the corresponding level of waste heat and user demand is generated. Among them, the adaptability of flow fluctuation is obtained by calculating the similarity between the waste heat source flow curve and the user demand baseline through a dynamic morphological matching algorithm. When the matching degree is lower than the preset threshold, the coupling path optimization module starts the multi-source collaborative coupling mode, that is, selects two waste heat sources at different levels, calculates the adjustable heat capacity according to the real-time parameters of each waste heat source, and synthesizes the target heat parameters according to the heat capacity ratio, so that the matching degree between the target heat parameters and the user demand characteristic parameters reaches above the preset threshold, and generates a collaborative coupling path that includes the output ratio of each waste heat source and the transmission path switching node.

2. The data-driven industrial circulating cooling water waste heat multi-source coupling system according to claim 1, characterized in that, The method for establishing dynamic characteristic profiles of waste heat at each level includes: A time series neural network is used to perform adaptive window feature extraction on the monitoring data. The prediction model is used to predict the parameter fluctuation trend. Real-time fluctuation data is dynamically compared with historical patterns and abnormal conditions are marked. At the same time, the probability characteristic envelope of the parameter change curve is generated.

3. The data-driven industrial circulating cooling water waste heat multi-source coupling system according to claim 1, characterized in that, The matching degree calculation method is as follows: The coupled path optimization module constructs a multi-dimensional feature vector space, uses a similarity algorithm to calculate the matching value between dynamic characteristic files and heat demand feature parameters and maps them to a standard interval. When multiple waste heat sources meet the threshold conditions, a multi-objective optimization screening mechanism is introduced to perform dual priority sorting and output a list of non-dominated solutions.

4. The data-driven industrial circulating cooling water waste heat multi-source coupling system according to claim 1, characterized in that, The temperature range overlap is defined as the proportion of overlap between the user's required temperature range and the temperature fluctuation range of the waste heat source. The flow fluctuation adaptability is calculated by using a dynamic morphological matching algorithm to determine the similarity between the waste heat source flow curve and the user's required baseline. The duration fit is the degree to which the sustainable heating duration of the waste heat source covers the duration of the user's required heating.

5. The data-driven industrial circulating cooling water waste heat multi-source coupling system according to claim 1, characterized in that, The logic for generating the coupling path is as follows: The intelligent coupling control unit (3) selects the waste heat source with the highest matching degree as the main power source based on the industrial scenario topology map, determines the transmission path with the lowest energy consumption through the path search algorithm, pre-adjusts the opening degree of the intelligent regulating valve group, and uses the digital twin platform to simulate and verify the feasibility of the path.

6. The data-driven industrial circulating cooling water waste heat multi-source coupling system according to claim 1, characterized in that, The strategy for selecting two waste heat sources at different levels is as follows: The coupling path optimization module excludes waste heat levels that deviate from the target temperature range by a preset margin, constructs a heat capacity contribution evaluation matrix in the remaining levels, and selects waste heat source combinations whose effective heat capacity meets the redundancy requirements of waste heat users and whose high-temperature level proportion reaches the minimum guaranteed value.

7. The data-driven industrial circulating cooling water waste heat multi-source coupling system according to claim 1, characterized in that, The method for calculating the adjustable heat capacity is as follows: The instantaneous heat capacity values ​​of each waste heat source are dynamically updated based on real-time collected temperature and flow data. At the same time, a pipeline attenuation compensation mechanism is introduced to pre-correct heat capacity loss during long-distance transmission.

8. The data-driven industrial circulating cooling water waste heat multi-source coupling system according to claim 1, characterized in that, The logic for the weighted synthesis of the target thermal parameters is as follows: Using the heat capacity ratio of different levels of waste heat sources as weighting coefficients, the temperature parameters are linearly weighted and fused, and the flow parameters are smoothed and fused, so that the fluctuation range of the synthesized target heat parameters is lower than the original fluctuation level of the single source.

9. The data-driven industrial circulating cooling water waste heat multi-source coupling system according to claim 1, characterized in that, The process of generating the cooperative coupling path is as follows: The flow distribution valve is driven by the heat capacity ratio. Pressure buffer devices are deployed at the transmission path switching nodes to prevent fluid impact. A control sequence containing multi-level response commands is generated to realize the time-sequential switching and parameter adjustment of the coupled path.

Images