Energy consumption digital management method and system based on energy management platform

Abstract

The invention belongs to the technical field of energy management, and provides an energy consumption digital management method and system based on an energy management platform. The method comprises the steps of comparing real-time energy consumption data with a dynamic expected reference value to obtain a theoretical static potential value when it is judged that an optimization management condition is met based on each piece of real-time energy consumption data, and introducing confidence interval judgment to correct the theoretical static potential value to obtain an optimization potential value with statistical significance; the optimization contribution value of each energy consumption unit is obtained through analysis according to the optimization potential value, the association topological relation and a preset energy consumption management target; and according to the optimization potential value and the optimization contribution value, screening to obtain a group of energy consumption units, and only processing and displaying the energy consumption data of the group of energy consumption units. According to the method, the existing system data piling presentation mode is changed, so that the core energy-saving opportunity can be quickly positioned from mass data, limited management resources are concentrated in a link which contributes to the global energy efficiency target to the greatest extent, and secondary optimization decision and resource dispersion are effectively avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy management, in particular to an energy consumption digital management method and system based on an energy management platform. BACKGROUND

[0002] The existing energy consumption digital management system usually deploys intelligent sensors, intelligent electricity meters and other data acquisition devices to monitor the consumption of water, electricity, gas and heat in real time, and uploads the data to a cloud energy management platform. The platform stores, aggregates and performs basic calculations on the data, and finally displays the overall energy consumption, sub-energy consumption, energy consumption trend, same period and comparative analysis results to the manager in the form of visualizations such as dashboards, line charts, column charts and pie charts.

[0003] However, the existing energy consumption digital management system tends to pursue a large and complete set of data, and piles up a large amount of energy consumption information on the manager interface without screening. The manager faces complex data charts and is difficult to quickly identify the most critical and effective energy-saving opportunity point, easily falling into the dilemma of having data but no insight, which seriously reduces the decision-making efficiency and makes the management measures often lag behind.

[0004] At the same time, the functions of the existing energy consumption digital management system are limited to monitoring and statistics, which leads to the fact that even if multiple energy consumption abnormal points are found, it is impossible to determine which device or link contributes the most to achieving the overall energy efficiency improvement goal (such as reducing total electricity cost or completing unit product energy consumption indicators). This leads to very scattered energy consumption management actions and cannot meet the needs of managers.

[0005] Therefore, there is an urgent need for a new type of energy consumption digital management method to shift from passive display to active decision support, thereby providing effective assistance for managers to carry out efficient and accurate energy management. SUMMARY

[0006] To solve the above technical problems, the present application provides an energy consumption digital management method and system based on an energy management platform.

[0007] The present application provides an energy consumption digital management method based on an energy management platform, comprising the following steps: Obtain real-time energy consumption data, working condition data of a plurality of energy-using units to which the target area belongs, and the associated topological relationship between the energy-using units, and determine whether the optimization management condition is met based on each real-time energy consumption data; When it is determined that the optimization management condition is met, the real-time energy consumption data is compared with a dynamic expected benchmark value to obtain a theoretical static potential value, a confidence interval determination is introduced to correct the theoretical static potential value, and an optimization potential value with statistical significance is obtained; wherein the dynamic expected benchmark value is obtained based on a dynamic energy consumption benchmark model and the working condition data, and the dynamic energy consumption benchmark model is constructed based on historical data; An optimization contribution value of each energy-using unit is obtained according to the optimization potential value, the associated topological relationship and a preset energy consumption management target; A group of energy-using units is screened according to the optimization potential value and the optimization contribution value, and only the energy consumption data of the group of energy-using units is processed and displayed.

[0008] The application further provides an energy consumption digital management system based on an energy management platform, and the system comprises: A triggering unit acquires real-time energy consumption data, working condition data of a plurality of energy-using units to which a target area belongs and an associated topological relationship between the energy-using units, and determines whether an optimization management condition is met based on each real-time energy consumption data; An optimization potential analysis unit compares the real-time energy consumption data with a dynamic expected benchmark value to obtain a theoretical static potential value when it is determined that the optimization management condition is met, introduces a confidence interval determination to correct the theoretical static potential value, and obtains an optimization potential value with statistical significance; wherein the dynamic expected benchmark value is obtained based on a dynamic energy consumption benchmark model and the working condition data, and the dynamic energy consumption benchmark model is constructed based on historical data; An optimization contribution analysis unit obtains an optimization contribution value of each energy-using unit according to the optimization potential value, the associated topological relationship and a preset energy consumption management target; A digital management unit screens a group of energy-using units according to the optimization potential value and the optimization contribution value, and only processes and displays the energy consumption data of the group of energy-using units.

[0009] The above scheme of the application can accurately identify energy-using units with significant energy-saving potential based on dynamic benchmark and statistical confidence analysis, and screen the most critical and effective optimization targets by evaluating the global contribution values of the energy-using units. Furthermore, only a small amount of high-value data is processed and displayed. The application changes the presentation mode of the existing system data stacking, so that managers can quickly locate core energy-saving opportunities from massive data, concentrate limited management resources on the link that contributes most to the global energy efficiency target, effectively avoid suboptimal decisions and resource dispersion, and thus improve the decision-making efficiency and execution accuracy of energy management, and realize the change from passive monitoring to active and accurate decision support. BRIEF DESCRIPTION OF DRAWINGS

[0010] Figure 1 is a flow diagram of an energy consumption digital management method based on an energy management platform according to an embodiment of the present application; Figure 2 is a scene diagram of an energy consumption digital management system based on an energy management platform according to an embodiment of the present application; Figure 3 is a structure diagram of an energy consumption digital management system based on an energy management platform according to an embodiment of the present application. DETAILED DESCRIPTION

[0011] The technical solutions of the present application will be further described in detail below with the aid of the accompanying drawings and embodiments.

[0012] To make the objectives, technical solutions and advantages of embodiments of the present application clearer, the technical solutions in embodiments of the present application will be described clearly and completely below with the aid of the accompanying drawings. Obviously, the described embodiments are only some but not all of the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of the present application.

[0013] To facilitate the understanding of embodiments of the present application, further explanation and description will be made below with specific embodiments in conjunction with the accompanying drawings. The embodiments do not constitute a limitation on embodiments of the present application.

[0014] As shown in Figure 1 , the present application discloses an energy consumption digital management method based on an energy management platform, comprising the following steps: 100, obtaining real-time energy consumption data, working condition data of a plurality of energy-using units to which a target area belongs, and associated topological relationships between the energy-using units, and determining whether an optimization management condition is satisfied based on each of the real-time energy consumption data; In this step, as shown in Figure 2 , the scheme of the present application is applied to an energy management platform, which is, for example, a cloud or a server. A plurality of energy-using units of each target area access the energy management platform with the region as an attribute. Then, the energy management platform obtains real-time energy consumption data, working condition data of a plurality of energy-using units to which a target area belongs, and associated topological relationships between the energy-using units. It can be understood that the target area can be a production workshop, a commercial building or an industrial park, etc., and the present application does not make a specific limitation thereon.

[0015] Based on the above system architecture, the energy management platform receives all real-time energy consumption data (such as power, current, and electricity) of energy consumption units (such as air compressors, refrigeration units, fans, lighting circuits, and production equipment) in a single target area, working condition data (such as device load rate, running time, ambient temperature, and planned output) of the energy consumption units, and associated topological relationships (such as process upstream and downstream relationships, energy flow directions, and membership in the same system) between the energy consumption units. Then, for example, based on whether the deviation of each real-time energy consumption data from the expected energy consumption benchmark value continuously exceeds a threshold value, it is determined whether the optimization management condition is met.

[0016] 200, when it is determined that the optimization management condition is met, comparing the real-time energy consumption data with a dynamic expected benchmark value to obtain a theoretical static potential value, introducing a confidence interval determination to correct the theoretical static potential value to obtain an optimization potential value with statistical significance; wherein the dynamic expected benchmark value is obtained based on a dynamic energy consumption benchmark line model and the working condition data, and the dynamic energy consumption benchmark line model is constructed based on historical data; In this step, when it is determined that the optimization management condition is met, it is determined that the energy consumption performance of the energy consumption units in the target area is poor or not as expected, and optimization processing is needed. That is, the special digital output management method of the present application, which only processes and displays the energy consumption data of specific groups of energy consumption units, needs to be started to assist managers to perform more efficient and accurate energy management.

[0017] Specifically, first, the real-time energy consumption data is compared with the dynamic expected benchmark value to calculate a preliminary theoretical static potential value. The dynamic expected benchmark value is not a fixed value, but a theoretical expected energy consumption value under the current working condition calculated based on a dynamic energy consumption benchmark line model constructed based on historical data and combined with the current working condition data (such as current output and ambient temperature). It can be understood that the dynamic energy consumption benchmark line model can reflect the expected energy consumption level of the energy consumption unit under specific production or environmental conditions.

[0018] To further improve the reliability of the analysis results, the present application introduces a confidence interval determination method based on a statistical process to correct the theoretical static potential value. Specifically, if the difference between the real-time energy consumption data and the dynamic expected benchmark value continuously exceeds the confidence interval (such as 95% or 99% confidence) calculated according to the historical fluctuation range, it is determined that the potential value is valid, and the difference is determined as the optimization potential value with statistical significance; if it does not exceed, it is determined to be a normal fluctuation, and its optimization potential value is set to zero or marked as a low priority.

[0019] 300, according to the optimization potential value, the associated topological relationship, and the preset energy consumption management target, an optimization contribution value of each energy consumption unit is analyzed and obtained; In this step, firstly, a linkage effect model is constructed based on the interconnected topology to simulate and calculate the chain reaction (i.e., side effects) in energy consumption of related units caused by reducing the energy consumption of a certain energy-consuming unit. For example, turning off a cooling fan may cause the ambient temperature to rise, leading to an increase in air conditioning energy consumption. Then, combined with a production performance model, the initial contribution of the energy-saving potential (optimization potential value) of this energy-consuming unit to the overall energy management goal is calculated. Finally, a system net benefit model is used to superimpose the initial contribution with the negative effects brought about by the chain reaction, and the resulting net value is the optimized contribution value of this energy-consuming unit.

[0020] This step ensures that priority ranking serves global energy efficiency optimization, rather than local optimization, thus preventing suboptimal problems.

[0021] 400. Based on the optimization potential value and the optimization contribution value, a group of energy-consuming units is selected, and only the energy consumption data of this group of energy-consuming units is processed and displayed.

[0022] In this step, based on the aforementioned optimization potential and optimization contribution values, each energy-consuming unit within the target area is screened to obtain a group of effective energy-consuming units. For example, the units are sorted according to their product or weighted sum, and the most critical and effective energy-consuming units with rankings above a predetermined threshold are selected. Then, only the energy consumption data of this group of units is processed in depth (e.g., root cause analysis, generating optimization suggestion reports) and presented to managers in a visual format (e.g., highlighted alarms, optimization suggestion pop-ups). In this way, managers are no longer faced with complex charts, but rather a list of high-value information that has been intelligently filtered and can directly support decision-making. This allows them to quickly identify core energy-saving opportunities, concentrate management resources on core energy-consuming units, and significantly improve decision-making efficiency and the accuracy of energy management measures. Under normal circumstances, managers can still be provided with an unfiltered, full-data presentation.

[0023] The above-mentioned solution of the present invention can accurately identify energy-consuming units with significant energy-saving potential based on dynamic benchmarks and statistical confidence analysis. By evaluating their global contribution value, it can screen out the most critical and effective optimization targets. Furthermore, it can perform in-depth processing and display only the selected small amount of high-value data. By changing the data-stacking presentation method of existing systems, the present invention enables managers to quickly locate core energy-saving opportunities from massive amounts of data, concentrating limited management resources on the links that contribute the most to the overall energy efficiency goals. This effectively avoids suboptimal decisions and resource dispersion, thereby improving the decision-making efficiency and execution accuracy of energy management, and realizing the transformation from passive monitoring to proactive and precise decision support.

[0024] As an example, determining whether the optimization management conditions are met based on the real-time energy consumption data includes: The total energy consumption value of the target area is calculated based on the real-time energy consumption data of all energy consumption units in the current statistical period, and the total energy consumption value is compared with a preset energy consumption threshold value; when the total energy consumption value continuously exceeds the energy consumption threshold value and reaches a first preset time length, it is determined that the optimization management condition is met; and / or, The deviation rate of the real-time energy consumption data of each energy consumption unit from the individual expected energy consumption reference value is calculated, and the number of energy consumption units whose deviation rate exceeds a preset deviation threshold value is counted; when the number exceeds a preset number threshold value and continuously reaches a second preset time length, it is determined that the optimization management condition is met.

[0025] The embodiment provides two optional and specific determination implementations of whether the optimization management condition is met.

[0026] For the first determination method: based on the real-time energy consumption data of all energy consumption units in the target area in the current statistical period (such as the past 15 minutes or 1 hour), the total energy consumption value of the area (such as total power consumption or total gas consumption) is calculated. The calculated total energy consumption value of the area is compared with a preset energy consumption threshold value of the area. The energy consumption threshold value of the area can be set according to historical data, budget indicators or energy efficiency targets. In order to avoid misjudgment caused by instantaneous fluctuations, the present application also sets a time duration condition, that is, only when the total energy consumption value of the area continuously exceeds the energy consumption threshold value of the area and the exceeding state reaches a first preset time length (for example, continuously for 30 minutes), it is finally determined that the optimization management condition is met. This method is suitable for quickly finding overall energy efficiency degradation or abnormality, for example, sudden increase in energy consumption of the entire workshop or building.

[0027] For the second determination method: first, for each energy consumption unit in the area, the deviation rate of the real-time energy consumption data from the individual expected energy consumption reference value is calculated. The individual expected energy consumption reference value is a personalized reference value based on the historical operation data or theoretical model of each unit. A preset deviation threshold value (such as ±10%) is set, and the number of energy consumption units whose deviation rate exceeds the preset deviation threshold value is counted. Similarly, in order to exclude random interference, the present application sets that only when the number of abnormal units exceeds a preset number threshold value (for example, 20% of the total number), and this state continuously reaches a second preset time length (for example, continuously for 15 minutes), the optimization management condition is determined to be met. This method is suitable for finding widespread and scattered energy efficiency problems, for example, the general energy consumption level rises due to equipment aging or failure of unified strategy.

[0028] Through the setting of the two determination modes, the application can flexibly and reliably perceive energy efficiency abnormalities from the two dimensions of surface and quantity, ensure that the time-consuming deep analysis is triggered only when necessary, thereby significantly improving the data processing efficiency and resource utilization while ensuring the system sensitivity.

[0029] As an example, the real-time energy consumption data is compared with the dynamic expected benchmark value to obtain a theoretical static potential value, and a confidence interval determination is introduced to correct the theoretical static potential value to obtain an optimized potential value with statistical significance, including: 201. Construct a dynamic energy consumption benchmark line model for each energy-using unit based on historical data, input the current real-time working condition data into the dynamic energy consumption benchmark line model to obtain a dynamic expected benchmark value under the current working condition; In this step, the dynamic energy consumption benchmark line model reflects the mathematical mapping relationship between energy consumption and working conditions (such as environmental temperature, device load rate, production output, etc.). The specific explanation is as follows: The dynamic energy consumption benchmark line model is constructed based on the historical data of each energy-using unit. The historical data includes historical energy consumption data and corresponding, synchronously collected historical working condition data in the same period (such as the past few weeks, months, or years). Working condition data refers to key state parameters that affect the energy consumption of energy-using units, for example, for a central air conditioning chiller unit, its working condition data can include environmental temperature, humidity, cold water outlet temperature set value, device load rate, etc.; for an injection molding machine, its working condition data can include production mode, cycle time, mold temperature, hydraulic system pressure, etc.

[0030] The dynamic energy consumption benchmark line model essentially establishes a mathematical mapping relationship between energy consumption and multi-dimensional working condition variables, and its function form can be represented as: E_expected=f(x1,x2,x3,...,xn). Where E_expected is the dynamic expected benchmark value, x1,x2,x3,...,xn are n working condition variables that affect energy consumption. The specific form of function f is obtained by training and fitting historical data through machine learning algorithms, which can include but are not limited to: multiple linear regression (Multiple Linear Regression), random forest (Random Forest), gradient boosting decision tree (Gradient Boosting Decision Tree, GBDT / XGBoost), neural networks (Neural Networks), etc., and specific details are not repeated.

[0031] In the training phase, the selected algorithm is trained using the cleaned historical data (working condition data as features, energy consumption data as labels) to obtain a trained model. This process is usually completed offline during system initialization or periodic updates.

[0032] In the application (inference) stage, the collected current real-time working condition data (i.e. the values of x1, x2, x3,..., xn at the current moment) is input into the trained dynamic energy consumption baseline model, and the dynamic energy consumption baseline model can output a personalized dynamic expected baseline value (E_expected) for the current specific operating condition. The dynamic expected baseline value represents the energy consumption level that the energy using unit should achieve under the current specific condition.

[0033] 202, calculate the absolute difference between the real-time energy consumption data and the dynamic expected baseline value as a theoretical static potential value; In this step, the dynamic expected baseline value obtained in step 201 is compared with the measured real-time energy consumption data, and the absolute difference between them is calculated. This difference is the theoretical static potential value, which preliminarily quantifies the gap between the current actual energy consumption and the ideal baseline, reflecting the potential energy saving space.

[0034] 203, based on the historical energy consumption data fluctuation characteristics, a dynamic confidence interval corresponding to a pre-set confidence level is calculated; In this step, in order to distinguish between real anomalies and normal fluctuations, the concept of statistical process control is introduced. Specifically, based on the historical energy consumption data fluctuation characteristics (such as standard deviation, variance) of the energy using unit, a dynamic confidence interval is calculated, which is a dynamic range that changes with the statistical characteristics of the data itself, used to judge whether the energy consumption deviation is within the normal fluctuation range. The width of the dynamic confidence interval is determined by the historical volatility, and the pre-set confidence level (such as 95% or 99%) defines the probability of the interval containing normal fluctuation data.

[0035] 204, judge whether the theoretical static potential value continuously exceeds the dynamic confidence interval and the duration reaches a third pre-set time length; if yes, the theoretical static potential value is determined as an optimization potential value with statistical significance; if not, the optimization potential value is set to zero or marked as low priority.

[0036] In this step, it is continuously monitored whether the theoretical static potential value (i.e. the energy consumption deviation) continuously exceeds the dynamic confidence interval calculated in step 203, and whether the duration of the exceeding state reaches a third pre-set time length (for example, 10 consecutive minutes). The duration condition is used to filter out short-term spike pulse interference. If yes, it is determined that the energy consumption deviation has statistical significance and is not random noise, so the theoretical static potential value is finally determined as an effective optimization potential value for subsequent analysis. If not, it is determined that the deviation belongs to the normal fluctuation range, and its optimization potential value is set to zero or marked as low priority, so as to avoid subsequent in-depth analysis of insignificant fluctuations and save computing resources.

[0037] As an example, the optimization contribution value of each energy-using unit is obtained by analyzing the optimization potential value, the correlation topological relationship and a preset energy consumption management target, comprising: 301. A global optimization objective function is determined according to the preset energy consumption management target, and the energy consumption management target comprises total electricity cost minimization or unit product energy consumption minimization. In this step, the preset abstract energy consumption management target is analyzed into a quantifiable and calculable global optimization objective function. The energy consumption management target is a macroscopic index such as total electricity cost minimization or unit product energy consumption minimization. When the energy consumption management target is total electricity cost minimization, the global optimization objective function needs to introduce a time-of-use electricity price model to convert energy consumption (kWh) into electricity cost (yuan); when the energy consumption management target is unit product energy consumption minimization, the objective function needs to establish the correlation between total energy consumption and total output.

[0038] 302. A system linkage effect model for representing the energy flow mutual influence relationship between energy-using units is constructed based on the correlation topological relationship; for each energy-using unit, the energy consumption chain change value of the correlated units caused by the energy consumption change of the energy-using unit is simulated and calculated through the system linkage effect model according to the optimization potential value of the energy-using unit. In this step, the system linkage effect model is constructed based on the correlation topological relationship (such as process sequence, energy flow direction and device dependency relationship) obtained in step 100, which defines the mutual influence relationship of energy flow between energy-using units.

[0039] The purpose of setting the system linkage effect model is to overcome the limitation of traditional energy consumption management which only focuses on a single device or isolated link, to evaluate the true value of energy-saving measures from a global perspective by quantifying the energy interaction and mutual influence between energy-using units, and to prevent suboptimal problems such as local energy saving and global energy consumption.

[0040] The system linkage effect model is constructed based on the correlation topological relationship obtained in step 100. The correlation topological relationship defines the connection and dependency relationship of energy flow, material flow or information flow between energy-using units in the target area, specifically including: (1) process sequence relationship: for example, devices A, B and C on the production line are arranged in the processing sequence, and the start-stop or energy consumption change of device A directly affects the production rhythm and energy consumption of subsequent devices B and C; (2) energy flow direction relationship: for example, the cold energy generated by the refrigeration unit is delivered to the terminal air conditioning box through the water pump, and the energy consumption change (such as frequency regulation) of the water pump will affect the water flow, and then affect the operating efficiency and energy consumption of the refrigeration unit; (3) environmental coupling relationship: for example, the heat dissipation of multiple devices collectively affects the environmental temperature of the workshop, and the environmental temperature in turn affects the load and energy consumption of the workshop air conditioning system; (4) belonging to the same system relationship: for example, multiple lighting loops jointly provide lighting for a region, and their on-off logic is linked.

[0041] The system linkage effect model is essentially to establish a directed graph network or system equation for representing and calculating the cascading changes in energy consumption of other associated nodes (result nodes) caused by changes in energy consumption of a node (cause node). The mathematical expression can be simplified as: ΔE_j=g(ΔE_i,R_ij,State) Where ΔE_i represents the change in energy consumption of the cause node (e.g. device i) (i.e. its optimization potential value); R_ij represents the influence relationship coefficient (e.g. conduction efficiency, conversion coefficient, heat exchange efficiency, etc.) between the cause node i and the result node j, which is learned or set from the associated topological relationship and historical operation data; State represents the current state of the system (e.g. current ambient temperature, current total load level); ΔE_j represents the cascading change value of the energy consumption of the result node j; g is a function describing this influence relationship, which can be a linear function, a nonlinear function, or a black box model based on machine learning (such as neural network), depending on the complexity of the system.

[0042] In modeling, according to the associated topological relationship, the connection relationship and direction of all nodes in the network are determined. Through historical data analysis and system mechanism, an initial influence relationship coefficient (R_ij) or influence function g is assigned or trained for each connection relationship.

[0043] In calculating the cascading change value of energy consumption, for a to-be-evaluated energy-using unit i, its optimization potential value (ΔE_i) is input into the model. The model automatically simulates and outputs the set of cascading change values (ΔE_j) of energy consumption of all associated units j caused by this. These change values can be negative (energy consumption reduction) or positive (energy consumption increase). For example, the quantitative value of the simulation of turning off a cooling fan (energy consumption reduction) leading to an increase in the temperature of the computer room, which in turn forces the precision air conditioner to increase the refrigeration power (energy consumption increase).

[0044] 303, according to the global optimization objective function, calculate the direct benefit value brought by the change in energy consumption of the energy-using unit, superimpose the direct benefit value and the cascading change value of energy consumption to obtain the net benefit value of the system, which is taken as the optimization contribution value of the energy-using unit; In this step, first, according to the global optimization objective function determined in step 301, the direct benefit value (such as saved electricity cost, reduced unit product energy consumption) brought by the optimization potential value of the energy-using unit is calculated. Then, the calculated direct benefit value is superimposed with the cascading change value of energy consumption (usually negative benefit) simulated in step 302. The net value obtained by superimposing the operation is the real and net benefit brought by the energy-saving action of the energy-using unit for the whole system, i.e. the system net benefit value, which is determined as the optimization contribution value of the energy-using unit.

[0045] 304, repeat the above steps until all energy-using units are traversed, and obtain the optimized contribution value set of each energy-using unit.

[0046] By traversing all energy-using units whose optimized potential values have been calculated, an optimized contribution value based on the global system consideration is calculated for each energy-using unit, and finally an optimized contribution value set is formed.

[0047] As an example, for each energy-using unit, based on its optimized potential value, the energy consumption change of the associated units triggered by the energy consumption change of the energy-using unit is simulated and calculated through the system linkage effect model, including: 3021, based on the optimized potential values of the energy-using units and real-time working condition data, a scene feature matrix is constructed, a current energy-using unit running scene is identified based on the scene feature matrix, and a random disturbance signal is generated based on historical fluctuation data corresponding to the energy-using unit running scene; In this step, a scene feature matrix is constructed based on the optimized potential values of the energy-using units and real-time working condition data. The scene feature matrix contains macro statistical features, such as the sum, mean, variance, distribution of the optimized potential values, and the coupling relationship with the working conditions. These macro statistical features comprehensively reflect the current overall optimization situation in the target area, for example: whether it is a high-efficiency stable scene with slight optimization space, a high-potential optimization scene with a large number of devices with serious energy efficiency degradation, or a key breakthrough scene with optimization potential concentrated in a few key devices.

[0048] Subsequently, based on the overall system state described by the scene feature matrix, the current energy-using unit running scene is identified by matching with the historical typical scene library. After accurately identifying the system-level running scene, a random disturbance signal is generated based on the historical fluctuation data under the specific running scene. The historical fluctuation data is preferably a statistical summary of the fluctuation characteristics of each unit in the target area under the same overall optimization scene in history. The generated random disturbance signal can thus represent the typical fluctuation patterns that are expected to occur under this type of overall system state, which can reflect the overall uncertainty characteristics of the region rather than the isolated noise of a single device.

[0049] 3022, superimpose the optimized potential value and the random disturbance signal to generate a plurality of simulated energy consumption change values with fluctuation characteristics, and input each simulated energy consumption change value into the system linkage effect model to respectively simulate and calculate a plurality of sets of energy consumption chain change values of the associated units; In this step, the random interference signal generated in step 3021 is superimposed with the optimization potential value (i.e., the expected baseline energy consumption change) to generate multiple (e.g., N=1000) simulated energy consumption changes with fluctuating characteristics. Each simulated quantity represents one possible change in the energy consumption of the energy-consuming unit after considering real-world uncertainties.

[0050] Subsequently, these N simulated energy consumption changes are sequentially input into the system linkage effect model. For each input scenario, the system linkage effect model calculates a set of corresponding energy consumption chain change values ​​for the associated units. Finally, through numerous repeated simulations, N sets of energy consumption chain change values ​​can be obtained. It can be understood that this simulation process is essentially a Monte Carlo simulation, approximating the system's true response distribution through extensive random sampling.

[0051] 3023. Calculate the expected value of multiple sets of energy consumption chain change values, and use it as the final simulation result of the energy consumption chain change value of the associated units triggered by the energy-consuming unit.

[0052] In this step, the multiple sets of energy consumption chain change values ​​obtained in the above steps are statistically analyzed, and their expected value (mean) is calculated. This represents the average expected level of energy consumption chain change of related units after considering typical fluctuations under the current overall optimizable scenario of the system, and is ultimately used as the final simulation result.

[0053] In this embodiment, a random interference signal is generated using historical fluctuation data corresponding to the operating scenarios of the energy-consuming units, achieving accurate simulation of system uncertainties. By extracting historical fluctuation characteristics under specific operating scenarios, the generated random signal can realistically reflect the typical fluctuation patterns under such scenarios, rather than simple white noise interference. This method can effectively simulate the complex fluctuation characteristics existing in actual systems, making the subsequent Monte Carlo simulation results closer to reality. Through a large number of random simulations with scenario characteristics, the final mathematical expectation value can smooth out the influence of random fluctuations, significantly improving the accuracy and reliability of predicting the cascading changes in energy consumption of related units.

[0054] like Figure 3 As shown, this embodiment of the invention also provides an energy consumption digital management system 200 based on an energy management platform, comprising: Triggering unit 2001 acquires real-time energy consumption data, operating condition data, and the associated topology relationship between multiple energy-consuming units belonging to the target area, and determines whether the optimization management conditions are met based on the real-time energy consumption data. The optimization potential analysis unit 2002 obtains a theoretical static potential value by comparing the real-time energy consumption data with a dynamic expected benchmark value when it is determined that the optimization management condition is met, corrects the theoretical static potential value by introducing a confidence interval determination, and obtains an optimization potential value with statistical significance; wherein the dynamic expected benchmark value is obtained based on a dynamic energy consumption baseline model and the working condition data, and the dynamic energy consumption baseline model is constructed based on historical data; The optimization contribution analysis unit 2003 analyzes the optimization contribution value of each energy-using unit according to the optimization potential value, the associated topological relationship, and a preset energy consumption management target; The digital management unit 2004 filters a group of energy-using units according to the optimization potential value and the optimization contribution value, and only processes and displays the energy consumption data of the group of energy-using units.

[0055] As an example, the triggering unit 2001 is specifically configured to: calculate a regional total energy consumption value based on the real-time energy consumption data of all energy-using units in a target region in a current statistical period, compare the regional total energy consumption value with a preset regional energy consumption threshold, and determine that the optimization management condition is met when the regional total energy consumption value continuously exceeds the regional energy consumption threshold and reaches a first preset time length. and / or, calculate the deviation rate of the real-time energy consumption data of each energy-using unit from its individual expected energy consumption benchmark value, count the number of energy-using units whose deviation rate exceeds a preset deviation threshold, and determine that the optimization management condition is met when the number exceeds a preset number threshold and continuously reaches a second preset time length.

[0056] As an example, the optimization potential analysis unit 2002 is specifically configured to: construct a dynamic energy consumption baseline model of each energy-using unit based on historical data, input the current real-time working condition data into the dynamic energy consumption baseline model, and obtain a dynamic expected benchmark value under the current working condition; calculate the absolute difference between the real-time energy consumption data and the dynamic expected benchmark value as a theoretical static potential value; calculate a dynamic confidence interval corresponding to a preset confidence level based on the fluctuation characteristics of historical energy consumption data; determine whether the theoretical static potential value continuously exceeds the dynamic confidence interval and the duration reaches a third preset time length; if yes, the theoretical static potential value is determined as an optimization potential value with statistical significance; if no, the optimization potential value is set to zero or marked as a low priority.

[0057] As an example, the optimization contribution analysis unit 2003 is specifically configured to: determine a global optimization objective function according to the preset energy consumption management objective, the energy consumption management objective including total electricity bill minimization or unit product energy consumption minimization; construct a system linkage effect model for representing energy flow mutual influence relationship between energy using units based on the associated topological relationship; for each energy using unit, simulate and calculate energy consumption chain change value of associated units triggered by energy consumption change of the energy using unit according to optimization potential value of the energy using unit through the system linkage effect model; calculate direct benefit value brought by energy consumption change of the energy using unit according to the global optimization objective function, superimpose the direct benefit value and the energy consumption chain change value to obtain system net benefit value, and take the system net benefit value as optimization contribution value of the energy using unit; repeat the above steps until all energy using units are traversed to obtain optimization contribution value set of each energy using unit.

[0058] As an example, the optimization contribution analysis unit 2003 is specifically configured to: construct a scene feature matrix based on optimization potential value of each energy using unit and real-time working condition data, identify current energy using unit running scene based on the scene feature matrix, and generate random interference signal based on historical fluctuation data corresponding to the energy using unit running scene; superimpose the optimization potential value and the random interference signal to generate a plurality of simulated energy consumption change amounts with fluctuation characteristics, input each simulated energy consumption change amount into the system linkage effect model in sequence, and respectively simulate and calculate a plurality of groups of energy consumption chain change values of associated units; calculate mathematical expectation value of the plurality of groups of energy consumption chain change values, and take the mathematical expectation value as final simulation result of the energy consumption chain change value of the associated units triggered by the energy using unit.

[0059] Although the present application is specifically shown and described with reference to the preferred embodiments, those skilled in the art will appreciate that various modifications in form and detail can be made without departing from the spirit and scope of the application. Accordingly, the disclosed application is to be considered as illustrative only and is limited only as specified in the appended claims.

Claims

1. A method for digital energy consumption management based on an energy management platform, characterized in that, Includes the following steps: The system acquires real-time energy consumption data, operating condition data, and the associated topology relationships between multiple energy-consuming units within the target area, and determines whether the optimization management conditions are met based on the real-time energy consumption data. When the optimization management conditions are met, the real-time energy consumption data is compared with the dynamic expected benchmark value to obtain the theoretical static potential value. The theoretical static potential value is then corrected by introducing a confidence interval to obtain a statistically significant optimization potential value. The dynamic expected benchmark value is derived based on the dynamic energy consumption benchmark model and the operating condition data. The dynamic energy consumption benchmark model is constructed based on historical data. The optimization contribution value of each energy-consuming unit is derived from the optimization potential value, the associated topology relationship, and the preset energy consumption management target. Based on the optimization potential value and the optimization contribution value, a group of energy-consuming units is selected, and only the energy consumption data of this group of energy-consuming units is processed and displayed.

2. The energy consumption digital management method based on an energy management platform according to claim 1, characterized in that: Determining whether the optimization management conditions are met based on the aforementioned real-time energy consumption data includes: The total energy consumption value of the region is calculated based on the real-time energy consumption data of all energy-consuming units in the target area during the current statistical period. The total energy consumption value of the region is compared with a preset regional energy consumption threshold. When the total energy consumption value of the region continuously exceeds the regional energy consumption threshold and reaches a first preset time, it is determined that the optimization management condition is met. And / or, Calculate the deviation rate between the real-time energy consumption data of each energy-consuming unit and its individual expected energy consumption benchmark value, and count the number of energy-consuming units whose deviation rate exceeds a preset deviation threshold; when the number exceeds a preset number threshold and continues to reach a second preset duration, it is determined that the optimization management condition is met.

3. The energy consumption digital management method based on an energy management platform according to claim 1, characterized in that: The theoretical static potential value is obtained by comparing the real-time energy consumption data with the dynamic expected benchmark value. A confidence interval determination is then used to correct the theoretical static potential value, resulting in a statistically significant optimized potential value, including: Based on historical data, a dynamic energy consumption baseline model is constructed for each energy-consuming unit. The current real-time operating data is then input into the dynamic energy consumption baseline model to obtain the dynamic expected baseline value under the current operating conditions. Calculate the absolute difference between the real-time energy consumption data and the dynamic expected benchmark value, and use it as the theoretical static potential value; Based on the fluctuation characteristics of historical energy consumption data, a dynamic confidence interval corresponding to a preset confidence level is calculated; Determine whether the theoretical static potential value continuously exceeds the dynamic confidence interval for a duration of a third preset time; if yes, then determine the theoretical static potential value as an optimization potential value with statistical significance; if no, then set the optimization potential value to zero or mark it as low priority.

4. The energy consumption digital management method based on an energy management platform according to claim 3, characterized in that: Based on the optimization potential value, the associated topology, and the preset energy consumption management target, the optimization contribution value of each energy-consuming unit is analyzed, including: A global optimization objective function is determined based on the preset energy consumption management objective, wherein the energy consumption management objective includes minimizing total electricity costs or minimizing energy consumption per unit of product; Based on the aforementioned topological relationship, a system linkage effect model is constructed to characterize the mutual influence of energy flow among energy-consuming units. For each energy-consuming unit, based on its optimization potential value, the energy consumption chain change value of the associated units caused by its energy consumption change is simulated and calculated through the system linkage effect model. Based on the global optimization objective function, the direct benefit value brought about by the energy consumption change of the energy-consuming unit is calculated. The direct benefit value is superimposed with the energy consumption chain change value to obtain the system net benefit value, which is used as the optimization contribution value of the energy-consuming unit. Repeat the above steps until all energy-consuming units have been traversed, and obtain the set of optimized contribution values ​​for each energy-consuming unit.

5. The energy consumption digital management method based on an energy management platform according to claim 4, characterized in that: For each energy-consuming unit, based on its optimization potential value, the energy consumption chain reaction value of related units caused by its energy consumption change is simulated and calculated using the system linkage effect model, including: A scenario feature matrix is ​​constructed based on the optimization potential value and real-time operating data of each energy-consuming unit. The current operating scenario of the energy-consuming unit is identified based on the scenario feature matrix. A random interference signal is generated based on the historical fluctuation data corresponding to the operating scenario of the energy-consuming unit. The optimization potential value is superimposed with the random interference signal to generate multiple simulated energy consumption changes with fluctuation characteristics. Each simulated energy consumption change is then input into the system linkage effect model to simulate and calculate the energy consumption chain change values ​​of multiple sets of related units. Calculate the expected value of multiple sets of energy consumption chain change values, and use it as the final simulation result of the energy consumption chain change value of the associated units triggered by the energy-consuming unit.

6. An energy consumption digital management system based on an energy management platform, characterized in that, The system includes: The triggering unit acquires real-time energy consumption data, operating condition data, and the associated topology between multiple energy-consuming units belonging to the target area, and determines whether the optimization management conditions are met based on the real-time energy consumption data. The optimization potential analysis unit, when determining that the optimization management conditions are met, compares the real-time energy consumption data with the dynamic expected benchmark value to obtain the theoretical static potential value, and introduces a confidence interval judgment to correct the theoretical static potential value to obtain an optimization potential value with statistical significance; wherein, the dynamic expected benchmark value is derived based on the dynamic energy consumption benchmark model and the operating condition data, and the dynamic energy consumption benchmark model is constructed based on historical data; The optimization contribution analysis unit analyzes and derives the optimization contribution value of each energy-consuming unit based on the optimization potential value, the associated topology relationship, and the preset energy consumption management target. The digital management unit selects a group of energy-consuming units based on the optimization potential value and the optimization contribution value, and only processes and displays the energy consumption data of this group of energy-consuming units.

7. The energy consumption digital management system based on an energy management platform according to claim 6, characterized in that: The triggering unit is specifically used for: The total energy consumption value of the region is calculated based on the real-time energy consumption data of all energy-consuming units in the target area during the current statistical period. The total energy consumption value of the region is compared with a preset regional energy consumption threshold. When the total energy consumption value of the region continuously exceeds the regional energy consumption threshold and reaches a first preset time, it is determined that the optimization management condition is met. And / or, Calculate the deviation rate between the real-time energy consumption data of each energy-consuming unit and its individual expected energy consumption benchmark value, and count the number of energy-consuming units whose deviation rate exceeds a preset deviation threshold; when the number exceeds a preset number threshold and continues to reach a second preset duration, it is determined that the optimization management condition is met.

8. The energy consumption digital management system based on an energy management platform according to claim 6, characterized in that: The optimization potential analysis unit is specifically used for: Based on historical data, a dynamic energy consumption baseline model is constructed for each energy-consuming unit. The current real-time operating data is then input into the dynamic energy consumption baseline model to obtain the dynamic expected baseline value under the current operating conditions. Calculate the absolute difference between the real-time energy consumption data and the dynamic expected benchmark value, and use it as the theoretical static potential value; Based on the fluctuation characteristics of historical energy consumption data, a dynamic confidence interval corresponding to a preset confidence level is calculated; Determine whether the theoretical static potential value continuously exceeds the dynamic confidence interval for a duration of a third preset time; if yes, then determine the theoretical static potential value as an optimization potential value with statistical significance; if no, then set the optimization potential value to zero or mark it as low priority.

9. The energy consumption digital management system based on an energy management platform according to claim 8, characterized in that: The optimization contribution analysis unit is specifically used for: A global optimization objective function is determined based on the preset energy consumption management objective, wherein the energy consumption management objective includes minimizing total electricity costs or minimizing energy consumption per unit of product; Based on the aforementioned topological relationship, a system linkage effect model is constructed to characterize the mutual influence of energy flow among energy-consuming units. For each energy-consuming unit, based on its optimization potential value, the energy consumption chain change value of the associated units caused by its energy consumption change is simulated and calculated through the system linkage effect model. Based on the global optimization objective function, the direct benefit value brought about by the energy consumption change of the energy-consuming unit is calculated. The direct benefit value is superimposed with the energy consumption chain change value to obtain the system net benefit value, which is used as the optimization contribution value of the energy-consuming unit. Repeat the above steps until all energy-consuming units have been traversed, and obtain the set of optimized contribution values ​​for each energy-consuming unit.

10. The energy consumption digital management system based on an energy management platform according to claim 9, characterized in that: The optimization contribution analysis unit is specifically used for: A scenario feature matrix is ​​constructed based on the optimization potential value of each energy-consuming unit and real-time operating condition data. The current operating scenario of the energy-consuming unit is identified based on the scenario feature matrix. Random interference signals are generated based on historical fluctuation data corresponding to the operating scenarios of the energy-consuming units. The optimization potential value is superimposed with the random interference signal to generate multiple simulated energy consumption changes with fluctuation characteristics. Each simulated energy consumption change is then input into the system linkage effect model to simulate and calculate the energy consumption chain change values ​​of multiple sets of related units. Calculate the expected value of multiple sets of energy consumption chain change values, and use it as the final simulation result of the energy consumption chain change value of the associated units triggered by the energy-consuming unit.

Images