Energy efficiency improvement method and system based on double-layer agent swarming control

Abstract

The invention discloses an energy efficiency improvement method and system based on double-layer agent swarming control. The method specifically comprises the steps of obtaining a sample based on a database; defining a time decision unit, extracting an input index and constructing an output index; constructing an energy efficiency evaluation model based on data envelope analysis, and outputting an efficiency value, input and output relaxation values and various output indexes; the enterprise energy efficiency state is judged, and if the enterprise energy efficiency state is in a weak effective state, the enterprise energy efficiency is adjusted by adopting a double-layer agent swarming control strategy; inputting the adjusted data into an energy efficiency evaluation model for evaluation; and if the output relaxation value is not reduced, updating the double-layer agent swarming control adjustment strategy, and returning. According to the method, energy flow relevance and constraint among key equipment are fully considered, and energy adjustment can be effectively and flexibly carried out.

Description

Technical Field

[0001] The present invention belongs to the field of power information processing, and more specifically, relates to an energy efficiency improvement method and system based on double-layer intelligent agent swarming control. Background Art

[0002] With the deteriorating environment and the continuous enhancement of the greenhouse effect, the development model of "high energy consumption, high pollution, and huge energy consumption" has become a common problem faced by countries around the world. Since high-energy-consuming enterprises are large energy consumers, their energy supply systems often integrate multiple energy sources such as wind energy, solar energy, natural gas, coal, and electricity. There are multi-energy coupling and topological correlation structures between their equipment. It is necessary to consider multi-energy complementary operation methods on the basis of ensuring the stable and reasonable energy supply of various types of equipment, so as to improve the overall energy efficiency of high-energy-consuming enterprises. At the same time, with the advancement of new energy supply construction, enterprises often consider the resource endowment of the region to introduce wind energy, solar energy and other energy supply equipment and energy storage equipment, so it is necessary to further consider the problem of promoting the consumption of new energy, formulate a multi-energy scheduling optimization strategy for time-varying tracking, and realize the high-efficiency operation of multi-energy alliances. Traditional energy control methods are difficult to solve the coordination and optimization problems between different energy sources, including challenges in collaborative control, energy complementarity and flow tracking between multiple energy sources. Therefore, it is urgent to carry out research on energy efficiency improvement technology solutions for high-energy-consuming enterprises.

[0003] At present, in terms of improving enterprise energy efficiency, a patent with publication number CN105204333A has proposed an energy consumption prediction method for improving the energy utilization rate of steel enterprises. It uses a neural network algorithm with error feedback correction learning to predict the energy consumption of each process of the enterprise, improve energy utilization rate, and reduce total production costs. There is also a patent with publication number CN105204333A that proposes an enterprise energy efficiency management system and its control method, which automatically collects relevant data of various equipment in the thermal system in various links such as energy production, transportation, conversion, distribution, and consumption, calculates economic indicators, connects the indicators to form a curve, and adjusts the equipment according to the rising and falling trend of the curve. However, the existing technologies have the following problems: (1) They all fail to consider that enterprise energy supply is a dynamic time-varying process. In particular, with the advancement of new energy supply construction, enterprises need to take into account multi-energy coupling regulation and promote new energy consumption at the same time, and form a time-varying tracking energy regulation strategy that meets the quantitative evaluation of multiple indicators. (2) Only one indicator is used to evaluate the enterprise, and the preference of the enterprise's multi-dimensional indicators is not fully considered. (3) Due to the complex relationship between multiple energy supplies in high-energy-consuming enterprises, the existing technology for overall energy consumption control is difficult and ineffective. Summary of the invention

[0004] In order to solve the problems in the prior art that energy adjustment of high-energy-consuming enterprises does not take time-varying processes into consideration, the correlation between multiple energy supplies is complex, and the overall energy regulation efficiency is low, the present invention provides an energy efficiency improvement method based on double-layer intelligent agent swarm control. First, data envelopment analysis is used to construct an energy efficiency evaluation model for high-energy-consuming units to achieve a comprehensive evaluation of multiple input indicators and multiple output indicators of enterprise energy efficiency; then, when the enterprise's time decision unit fails to achieve optimal efficiency, a double-layer intelligent agent swarm control strategy is adopted according to the enterprise's time-varying supply capacity, combined with graph correlation and energy consumption constraints, to regulate the enterprise's energy input and track energy changes, and obtain the energy consumption input of each device and each energy at the next moment, thereby improving the energy efficiency of high-energy-consuming enterprises.

[0005] The present invention adopts the following technical solution.

[0006] The first aspect of the present invention proposes an energy efficiency improvement method based on double-layer intelligent agent swarming control, which is characterized by comprising the following contents:

[0007] Obtaining time-series energy usage samples of high-energy-consuming enterprises, wherein the time-series energy usage samples of high-energy-consuming enterprises include energy usage characteristic data, enterprise-related data, and energy constraint data;

[0008] Based on the energy consumption characteristic data, construct energy efficiency output indicators, wherein the energy efficiency output indicators include the enterprise energy saving cost ratio, comprehensive energy utilization rate and energy quality satisfaction;

[0009] Based on the energy consumption characteristic data, extract the different energy consumption data of the enterprise every hour as the input index, build an energy efficiency evaluation model for high-energy-consuming energy-consuming units based on data envelopment analysis, and output the efficiency value, input relaxation value, output relaxation value and energy efficiency output index of the enterprise's time series decision unit;

[0010] The energy efficiency status of the enterprise is judged according to the input relaxation value and the output relaxation value, and the energy efficiency evaluation threshold is set. If the input relaxation value and the output relaxation value are both less than or equal to the energy efficiency evaluation threshold, the energy efficiency unit time decision unit of the enterprise is determined to be valid at the current moment; if the input relaxation value or the output relaxation value is greater than the set energy efficiency evaluation threshold, the energy efficiency unit time decision unit of the enterprise is determined to be in a weakly valid state at the current moment;

[0011] When it is determined that the energy efficiency unit time decision unit of the enterprise is in a weak effective state at the current moment, a two-layer intelligent agent swarming control strategy is used to adjust the enterprise energy efficiency based on the enterprise association data and energy constraint data to obtain the different energy consumption inputs of each device at the next moment;

[0012] Based on the different energy consumption inputs of each device at the next moment, adjust the energy consumption inputs of each device at the current moment, input the adjusted data into the energy efficiency evaluation model of high-energy-consuming units, and obtain the input relaxation value, output relaxation value and various energy efficiency output indicators;

[0013] If the output slack value decreases, the energy consumption at the next moment will be adjusted based on the different energy consumption inputs of each device at the next moment; if the output slack value does not decrease, the two-layer intelligent agent swarm control adjustment strategy will be updated, and the two-layer intelligent agent swarm control strategy will be used to adjust the enterprise energy efficiency based on the enterprise-related data and energy constraint data.

[0014] Preferably, the energy consumption characteristic data includes the energy consumption cost, energy consumption and energy power of electricity, gas, heat, cold and new energy of enterprise equipment;

[0015] The enterprise-related data in the time-series energy consumption sample of the high-energy-consuming enterprise includes enterprise equipment physical topology data, enterprise equipment production-related data, and energy consumption-related data;

[0016] The energy constraint data in the time-series energy consumption samples of the high-energy-consuming enterprises include the enterprise energy constraint range, the energy output power constraint range at a moment, and the minimum and maximum values ​​of the total ramp rate of different energy supplies for different equipment at a moment.

[0017] Preferably, the enterprise energy-saving cost ratio in the energy efficiency output index is used to characterize the change in the cost ratio of electricity and gas charges at all times before and after the energy consumption regulation of the enterprise within the set time period and the electricity and gas charges at the corresponding time before the adjustment;

[0018] The comprehensive energy utilization rate in the energy efficiency output index represents the degree of complementarity between different energy sources of the enterprise within a set time period;

[0019] The energy quality satisfaction in the energy efficiency output index ensures the energy consumption efficiency of the enterprise.

[0020] Preferably, the calculation formula for the enterprise energy saving cost ratio in the energy efficiency output index is specifically:

[0021]

[0022] In the formula, C represents the energy-saving cost ratio of the enterprise; Respectively represent the energy consumption of the enterprise before and after the adjustment of the t s The change in electricity and gas charges at the data sampling time, among which, They represent the electricity and gas expenditure costs of the enterprise before the energy adjustment. They represent the electricity and gas expenditure costs of the enterprise after the energy consumption is adjusted; e represents electricity, g represents gas, and T represents a regulation time period, that is, the time period from the weak effective state before the adjustment to the end of the regulation. The T time is divided into several sampling moments, t s Indicates the data sampling time; the smaller C is, the more energy the enterprise saves and the lower the operating cost.

[0023] Preferably, the comprehensive energy utilization rate calculation formula in the energy efficiency output index is specifically:

[0024]

[0025] Where Rc represents the comprehensive energy utilization rate; r t′ Indicates any wind, solar, hydropower, thermal power, gas from t s Time to t s The complementary coefficient at the +1 moment, the smaller the complementary coefficient, the more the fluctuation of energy consumption is offset, and the stronger the complementarity between energy sources; They represent the cumulative energy consumption fluctuation rates of wind power, photovoltaic power, hydropower, thermal power, and gas in the time window t′ respectively; Respectively represent Energy in t s +1 time and t s Energy consumption at any time, For the Energy in [t s ,t s +1] Maximum energy consumption during the period, or gas.

[0026] Preferably, the energy quality satisfaction calculation formula in the energy efficiency output index is specifically:

[0027]

[0028] Where η ex Energy consumption for enterprises Efficiency; e , gas , h , c , rn They are the energy quality coefficient weights of electricity, gas, heat, cold, and new energy, which are equal to the proportion of each energy supply cost in the historical operation of the enterprise. The sum of the energy quality coefficient weights is 1; Respectively expressed in t s The electrical load power, heating load power and cooling load power of the nth device at the moment represent the effective energy value of the device; Respectively expressed in t sThe electricity, natural gas, thermal energy and new energy power input by the nth device at the moment, including solar energy input and wind energy input Same as equipment input new energy N represents the total number of devices.

[0029] Preferably, an energy efficiency evaluation model for high energy consumption units based on data envelopment analysis is constructed. DEA (λ,θ) specifically includes:

[0030]

[0031] Where x lk represents the lth input indicator data of the kth decision-making unit, where κ=1,2,…,K, K represents the total number of energy efficiency time decision-making units of the enterprise, l=1,2,…L, L represents the total number of input indicators; y rκ represents the energy efficiency output index data of the rth decision-making unit of the κth decision-making unit, where r = 1, 2, …, R, R represents the total number of output indicators; θ represents the efficiency value of the time decision-making unit; S - , S + Respectively represent the input relaxation index value and output relaxation index value; ε is 10 -5 ; l Represents the contribution weight of the lth input indicator to energy efficiency evaluation.

[0032] Preferably, the use of a two-layer intelligent agent swarming control strategy to adjust enterprise energy efficiency includes: with the goal of maximizing the energy efficiency output index of high-energy-consuming enterprises, setting a second intelligent agent responsible for actively tracking the changes in different energy inputs, and continuously and dynamically updating the coordinated inputs between different energy sources based on the changes; setting a first intelligent agent to determine the energy consumption inputs of different devices and different energy sources at the next moment based on the total energy consumption per unit time of a certain type of energy provided by the second intelligent agent, combined with the device interaction constraints associated with the graph, and form a strategy for increasing or decreasing the input of each energy of each device at the next moment.

[0033] Preferably, the specific scheme for adjusting the energy efficiency of an enterprise by adopting a double-layer intelligent agent swarming control strategy based on enterprise association data and energy constraint data is:

[0034] Build equipment interaction constraints based on the acquired enterprise-related data;

[0035] Construct the energy consumption change control equation of the two-layer intelligent agent, set nodes and expected goals;

[0036] Based on the acquired energy constraint data, a control input form of the second intelligent agent is formulated, and the second intelligent agent determines the accumulated energy consumption of different energy sources at the next moment according to the control input form;

[0037] In combination with the constructed device interaction constraints, a control input form of the first intelligent agent is formulated, and the first intelligent agent determines the energy consumption input of different energy sources and different devices at the next moment based on the control input form.

[0038] Preferably, the device interaction constraints are constructed in the following specific manner:

[0039] In the first agent's swarming control strategy, the equipment is used as an agent, and an undirected graph is used to represent the interaction between agents:

[0040] G(t)=(I,E(t))

[0041] Wherein, G(t) represents the interaction constraint graph between production equipment at time t, where t is the control time; I represents the set of equipment nodes, I = {1, 2, …, N}, and E(t) represents the set of edges between nodes at time t. If there is an interaction constraint between node i and node j at time t, there is a connected edge between the two nodes, and its value is (a, b), where i and j represent the i-th and j-th equipment nodes respectively, i∈[1, N], j∈[1, N], and N represents the total number of equipment; a represents the delay time for the downstream node to adjust the energy consumption after the signal of the upstream production node is sent to the downstream node, and its value is obtained according to the statistics of the production completion time of the upstream node; b represents the ratio of the year-on-year increase or decrease of the upstream and downstream nodes when adjusting the energy consumption, and its value is obtained according to the statistical data of various energy consumption statistics of the upstream and downstream nodes.

[0042] Preferably, the energy consumption change, nodes and expected goals of the two-layer agent are specifically:

[0043] Energy consumption change control equation:

[0044] The total number of the first agent and the second agent is Each agent is regarded as a node; its energy consumption changes conform to the following control equation:

[0045]

[0046] Among them, u α represents the control input item of the αth agent, as the unit energy consumption acceleration; v α represents the unit energy consumption of the αth following agent as the energy consumption rate; q α Represents the cumulative energy consumption of the αth following agent

[0047] node:

[0048] The first intelligent agent node is set as the device node, and the input is the energy consumption per unit time of each device with the same energy source; the second intelligent agent node is set as the node of all devices with different energy sources, and the input is the total energy consumption per unit time of all devices with different energy sources;

[0049] Expected goals:

[0050] The expected goal of the second intelligent agent is to maximize the comprehensive energy efficiency output index; the expected goal of the first intelligent agent is the total energy consumption per unit time of all devices of a certain energy source determined by the second intelligent agent.

[0051] Preferably, the second agent control input form is specifically:

[0052]

[0053] In the formula, They represent the cumulative energy consumption of the kth energy source at the next moment and this moment respectively. The cumulative amount of each energy source at the initial moment is a set random value; represents the unit energy consumption of the kth energy at the next moment and this moment, respectively, b k represents the energy input controller of the second agent, ε is the set energy efficiency evaluation threshold, which is 10 -5 ; Indicates the increase or decrease in the energy input of the second agent at the next moment; k represents the index allocation correction factor of the k-th energy source, K′ is the total number of energy categories, χ k , τ k They represent the energy cost of the k-th type of energy, comprehensive energy utilization rate and energy quality satisfaction of the enterprise, represents the weight of the contributing factors to the above three indicators, and The energy distribution ratio represents the comprehensive contribution value of the kth energy minus the average contribution value of all energy sources;

[0054] The adjusted energy amount of the second intelligent body must be within the constraints. The value of must be no less than 0 and cannot exceed the output power of the kth energy source at a certain moment. The value of must be between the maximum and minimum values ​​of the total ramp rate of all devices of the k-th energy source within a moment. If any value exceeds its range, the nearest boundary constraint value is taken as its value.

[0055] Preferably, the first agent control input form is specifically:

[0056]

[0057] Where Δf i g is the energy consumption control rule function f i gThe derivative of (v) represents the energy consumption adjustment per unit time between agents under the graph-related interaction constraints. The superscript g indicates that the function is an energy consumption control rule function. If there is an interaction constraint between two devices at time t, then their energy consumption control rule f i g (v) as follows:

[0058]

[0059] in, They represent the unit energy consumption of the jth device node at time t+1 and time t respectively. They represent the unit energy consumption of the i-th device node at time t+1 and time t respectively; They represent the optimal unit energy consumption of equipment j and equipment i respectively, which are the optimal unit energy consumption set by the factory parameters of the equipment or the optimal unit energy consumption obtained from daily operation statistics; It represents the optimal unit energy consumption ratio of device j and device i with interaction constraints; It represents the unit energy consumption ratio of device j and device i at time t with interaction constraints;

[0060] In the formula, f i γ It is the target feedback item, which is used to track the expected target and calculate the difference between the expected target and the total accumulated energy consumption of each device node. The superscript γ indicates that the function is the target feedback item function. The energy consumption increase or decrease of each device is determined according to the difference and the weighted proportion of the energy consumption of each device. The specific change control equation is:

[0061]

[0062] in, represents the cumulative energy consumption of the k-th second agent, whose value is the expected goal of the first agent, and k represents the k-th type of energy; represents the unit energy consumption of the i-th first agent, is the sum of the unit energy consumption of M first agents, where M is the number of first agents in the kth energy category; the energy consumption of the first agent following the second agent, and the absolute value of the difference is within the set threshold ε′, which is 0.1% of the energy consumption of the second agent, indicating that the first agent has completed tracking the energy consumption change of the second agent, otherwise, continue to use the above change control equation to increase or decrease the unit energy consumption of each first agent in proportion to complete the change tracking;

[0063] The equipment must meet the energy consumption constraint, and the adjusted amount of the first agent's energy must be greater than or equal to 0 and less than the maximum power limit of its input energy.

[0064] Preferably, the specific scheme for updating the double-layer agent swarming control adjustment strategy is: updating the random value set for the cumulative amount of each energy source at the initial moment when the second agent tracks the enterprise energy.

[0065] The second aspect of the present invention proposes an energy efficiency improvement system based on double-layer intelligent agent swarming control, comprising:

[0066] Sample collection module: used to obtain time-series energy consumption samples of high-energy-consuming enterprises, wherein the time-series energy consumption samples of high-energy-consuming enterprises include energy consumption characteristic data, enterprise-related data, and energy constraint data;

[0067] Energy efficiency output index construction module: used to construct energy efficiency output indicators based on the energy consumption characteristic data, wherein the energy efficiency output indicators include the enterprise energy saving cost ratio, comprehensive energy utilization rate and energy quality satisfaction;

[0068] Index output module: used to extract the different energy consumption data of the enterprise every hour based on the energy consumption characteristic data as the input index, build an energy efficiency evaluation model for high-energy-consuming energy-consuming units based on data envelopment analysis, and output the efficiency value, input relaxation value, output relaxation value and energy efficiency output index of the enterprise's time series decision unit;

[0069] Enterprise energy efficiency status judgment module: used to judge the enterprise energy efficiency status according to the input relaxation value and the output relaxation value, set the energy efficiency evaluation threshold, if the input relaxation value and the output relaxation value are both less than or equal to the energy efficiency evaluation threshold, then it is determined that the enterprise energy efficiency unit time decision unit is valid at the current moment; if the input relaxation value or the output relaxation value is greater than the set energy efficiency evaluation threshold, then it is determined that the enterprise energy efficiency unit time decision unit is in a weakly valid state at the current moment;

[0070] Enterprise energy efficiency adjustment module: if it is determined that the enterprise energy efficiency unit time decision unit is in a weak effective state at the current moment, it uses a two-layer intelligent agent swarming control strategy to adjust the enterprise energy efficiency based on the enterprise association data and energy constraint data, and obtains the different energy consumption inputs of each device at the next moment;

[0071] Enterprise energy efficiency adjustment progress judgment module: Based on the different energy consumption inputs of each device at the next moment, the energy consumption inputs of each device at the current moment are adjusted, and the adjusted data are input into the energy efficiency evaluation model of high-energy-consuming units to obtain input relaxation values, output relaxation values ​​and various energy efficiency output indicators; if the output relaxation value decreases, the energy consumption at the next moment will be adjusted based on the different energy consumption inputs of each device at the next moment; if the output relaxation value does not decrease, the two-layer intelligent agent swarm control adjustment strategy is updated, and the two-layer intelligent agent swarm control strategy based on enterprise-related data and energy constraint data is returned to adjust the enterprise energy efficiency.

[0072] The beneficial effect of the present invention is that, compared with the prior art,

[0073] (1) The present invention adopts graph association to set the interaction constraints between devices, and quantifies the interaction change constraints of energy flow between process-related devices by process delay and year-on-year rate increase or decrease. It fully considers the spatiotemporal characteristics and coupling characteristics of multiple energy flows such as cold, heat, electricity, and steam in the energy supply system of high-energy-consuming enterprises under different energy consumption boundaries, as well as the energy flow correlation and constraints between key equipment, ensuring that the energy supply control of various types of equipment is within a stable and reasonable range;

[0074] (2) The present invention adopts data envelopment analysis that meets multiple input-multiple output indicators to construct an energy efficiency evaluation model for high-energy-consuming energy-consuming units, takes time as the energy efficiency evaluation unit, selects the energy cost of the enterprise, the comprehensive energy utilization rate and the energy quality satisfaction of the enterprise for comprehensive evaluation of the energy efficiency of the enterprise, fully considers the preference of the multi-dimensional indicators of the enterprise subject, and can effectively improve the comprehensive energy efficiency of the time unit operation of high-energy-consuming enterprises by reducing the input relaxation value and the output relaxation value;

[0075] (3) The present invention adopts a two-layer intelligent agent swarm control to formulate an energy consumption control plan. The upper layer (i.e., the second intelligent agent) is responsible for determining the optimal coordinated input of various energy sources, and the lower layer (i.e., the first intelligent agent) determines a certain energy consumption input of each device based on a certain energy input determined by the upper layer. This solves the problem of complex multi-energy supply correlation and high difficulty in overall energy consumption control in high-energy-consuming enterprises. Through two-layer optimization scheduling, energy balance and optimal utilization are achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 An overall flow chart of the energy efficiency improvement method based on double-layer intelligent agent swarming control provided in an embodiment of the present disclosure;

[0077] Figure 2 A diagram of a double-layer intelligent agent swarming control structure provided in an embodiment of the present disclosure;

[0078] Figure 3 The control amount of the new energy supply input of each device before optimization of scenario 1 provided in the embodiment of the present disclosure;

[0079] Figure 4 The input control amount of new energy supply of each device after optimization of scenario 1 provided in the embodiment of the present disclosure;

[0080] Figure 5 A comparison chart of enterprise energy efficiency optimization effects in various typical scenarios provided in the embodiments of the present disclosure. DETAILED DESCRIPTION

[0081] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in this application are only embodiments of a part of the present invention, rather than all embodiments. Based on the spirit of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the protection scope of the present invention.

[0082] It should be noted that the following detailed descriptions are all illustrative and are intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present disclosure belongs.

[0083] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0084] like Figure 1 As shown, embodiment 1 of the present invention provides an energy efficiency improvement method based on double-layer intelligent agent swarming control, comprising:

[0085] Obtain time-series energy usage samples of high-energy-consuming enterprises, wherein the time-series energy usage samples of high-energy-consuming enterprises include energy usage characteristic data, enterprise-related data, and energy constraint data.

[0086] Preferably, the energy consumption characteristic data in the time series energy consumption samples of high energy consumption enterprises include the energy consumption cost, energy consumption and energy power of electricity, gas, heat, cold and new energy of enterprise equipment;

[0087] The enterprise-related data in the time-series energy consumption samples of high-energy-consuming enterprises include enterprise equipment physical topology data, enterprise equipment production-related data, and energy consumption-related data;

[0088] The energy constraint data in the time-series energy consumption samples of high-energy-consuming enterprises include the enterprise energy constraint range, the energy output power constraint range at a certain time, and the minimum and maximum values ​​of the total ramp rate of different energy supplies for different equipment at a certain time.

[0089] Specifically, the embodiment uses the new generation of electricity consumption information collection system of State Grid Corporation as the background platform, and obtains the time-series energy consumption sample data of A Steel Company from the electricity consumption information collection system of a certain province of State Grid, the energy metering integrated collection terminal, etc., including electricity, gas, heat, cold, new energy and other energy consumption characteristic data, enterprise-related data, energy constraint data, and selects historical data from January 2023 to January 2024 as the sample set. Among them, the power generation of photovoltaic and wind turbines has significant volatility according to the changes in the external environment; the thermal load is mainly used for production heat demand, and the demand fluctuation only changes with the production and operation mode of the enterprise, and is basically not affected by the external environment; industrial cold is mainly used for production cold and constant production environment temperature and other needs. The cold load shows a small fluctuation throughout the year, but the demand is large in summer.

[0090] Among them, Steel Company A has built energy conversion equipment such as photovoltaics, fans, electric energy storage, thermal energy storage, transformers, gas turbines, heat exchangers, and compression refrigerators to provide energy for production equipment such as rolling mills, rolling mills, heating furnaces, vertical rollers, conveyor rollers, hydraulic cylinders, coilers, straighteners, and flatteners. The sample includes the enterprise-related data formed by the above equipment.

[0091] In order to compare and illustrate the feasibility of this patented method for comprehensive energy efficiency evaluation and improvement of high-energy-consuming enterprises under changes in external environment, energy supply demand, and operation mode, a typical scenario of energy efficiency evaluation of steel enterprise A considering differences in energy demand and environmental changes is constructed, as shown in Table 1:

[0092] Table 1A Typical scenarios for energy efficiency assessment of steel enterprises

[0093]

[0094]

[0095] Among them, considering the volatility of photovoltaic and wind turbine power generation with changes in the external environment, by setting power generation thresholds respectively, two types of typical energy supply states were constructed, namely sufficient power generation and insufficient power generation; considering the large demand for cooling load in summer, two types of typical energy demand states were constructed, namely summer (25 degrees and above) and other seasons. A total of 8 typical scenarios were constructed and enterprise energy efficiency evaluation was carried out.

[0096] Based on the acquired energy consumption characteristic data, energy efficiency output indicators are constructed, wherein the energy efficiency output indicators include the enterprise energy saving cost ratio, comprehensive energy utilization rate and energy consumption quality satisfaction;

[0097] Preferably, the enterprise energy-saving cost ratio in the energy efficiency output index is used to characterize the change in the cost ratio of electricity and gas charges at all times before and after the energy consumption regulation of the enterprise within the set time period and the electricity and gas charges at the corresponding time before the adjustment;

[0098] The comprehensive energy utilization rate in the energy efficiency output index represents the degree of complementarity between different energy sources of the enterprise within a set time period;

[0099] The energy quality satisfaction in the energy efficiency output index ensures the energy consumption efficiency of the enterprise.

[0100] Preferably, the enterprise energy-saving cost index is comparative data. When steel enterprise A conducts energy efficiency regulation according to the strategy of increasing or decreasing the input of each equipment and energy at the next moment, the change of its energy consumption cost payment will directly affect the enterprise's operating costs. Therefore, the calculation formula of the enterprise energy-saving cost ratio is defined as:

[0101] The specific calculation formula for the enterprise energy saving cost ratio is:

[0102]

[0103] In the formula, C represents the energy-saving cost ratio of the enterprise; Respectively represent the energy consumption of the enterprise before and after the adjustment of the t s The change in electricity and gas charges at the data sampling time, among which,

[0104] They represent the electricity and gas expenditure costs of the enterprise before the energy adjustment. They represent the electricity and gas expenditure costs of the enterprise after the energy consumption is adjusted; e represents electricity, g represents gas, and T represents a regulation time period, that is, the time period from the weak effective state before the adjustment to the end of the regulation. The T time is divided into several sampling moments, t s Indicates the data sampling time; the smaller C is, the more energy the enterprise saves and the lower the operating cost.

[0105] Preferably, the comprehensive energy utilization rate calculation formula in the energy efficiency output index is specifically:

[0106]

[0107] Where Rc represents the comprehensive energy utilization rate; r t′ Indicates any wind, solar, hydropower, thermal power, gas from t s Time to t s The complementary coefficient at the +1 moment, the smaller the complementary coefficient, the more the fluctuation of energy consumption is offset, and the stronger the complementarity between energy sources; They represent the cumulative energy consumption fluctuation rates of wind power, photovoltaic power, hydropower, thermal power, and gas in the time window t′ respectively; Respectively represent Energy in t s +1 time and t s Energy consumption at any time, For the Energy in [t s ,t s +1] Maximum energy consumption during the period, or gas.

[0108] It should be noted that The smaller it is, the more the fluctuation of energy consumption of the multi-energy alliance is offset, and the stronger the complementarity between the energy sources. By quantitatively evaluating the complementarity of wind, light, water, fire, gas and other energy outputs as an energy efficiency evaluation indicator, the rationality of the coordinated operation model between energy sources can be improved based on the feedback adjustment of this indicator in the later stage, thereby improving the overall operational energy efficiency of A Steel Company;

[0109] Preferably, in view of the multi-energy operation of steel enterprise A, the stability of its energy supply should be ensured first, so as to avoid problems such as randomness of new energy output, which may cause losses to the enterprise's equipment and its production operations. The traditional energy analysis method does not take into account the changes in energy quality, so it is adopted The analytical method is used to evaluate the energy quality loss and use it as an energy efficiency evaluation indicator for energy quality satisfaction. The specific calculation formula for energy quality satisfaction is:

[0110]

[0111] Where η ex Energy consumption for enterprises Efficiency; e , gas , h , c , rn They are the energy quality coefficient weights of electricity, gas, heat, cold, and new energy, which are equal to the proportion of each energy supply cost in the historical operation of the enterprise. The sum of the energy quality coefficient weights is 1; They represent the electric load power, heat load power and cooling load power of the nth device at time ts, respectively, and represent the effective energy consumption value of the device; They represent the electric energy, natural gas, thermal energy and new energy power input to the nth device at time ts, among which the solar energy input and wind energy input Same as equipment input new energy N represents the total number of devices.

[0112] Based on the acquired energy consumption characteristic data, the different hourly energy consumption data of the enterprises contained therein are extracted as input indicators, and an energy efficiency evaluation model for high-energy-consuming units based on data envelopment analysis is constructed to output the efficiency value, input slack value, output slack value and energy efficiency output indicator of the enterprise's time series decision unit.

[0113] It should be noted that the enterprise timing decision unit is a decision on the hourly energy input of the enterprise.

[0114] Preferably, an energy efficiency evaluation model for high energy consumption units based on data envelopment analysis is constructed. DEA (λ,θ) specifically includes:

[0115]

[0116] Where x lk represents the lth input indicator data of the kth decision-making unit, where l = 1, 2, ..., K, K represents the total number of energy efficiency time decision-making units of the enterprise, l = 1, 2, ... L, L represents the total number of input indicators; y rk represents the energy efficiency output index data of the rth decision-making unit of the lth decision-making unit, where r = 1, 2, ..., R, R represents the total number of output indicators; θ represents the efficiency value of the time decision-making unit; S - , S + Respectively represent the input relaxation index value and output relaxation index value; ε is 10 -5 ; l Represents the contribution weight of the lth input indicator to energy efficiency evaluation.

[0117] The energy efficiency status of the enterprise is judged according to the input relaxation value and the output relaxation value, and the energy efficiency evaluation threshold is set. If the input relaxation value or the output relaxation value is greater than the set energy efficiency evaluation threshold, it is determined that the energy efficiency unit time decision unit of the enterprise is in a weakly effective state at the current moment, and the enterprise energy input index needs to be adjusted; if the input relaxation value and the output relaxation value are both less than or equal to the energy efficiency evaluation threshold, it is determined that the energy efficiency unit time decision unit of the enterprise is effective at the current moment.

[0118] Preferably, the energy efficiency evaluation threshold is 10 -5 .

[0119] When it is determined that the energy efficiency unit time decision unit of the enterprise is in a weakly effective state at the current moment, a two-layer intelligent agent swarming control strategy is used to adjust the energy efficiency of the enterprise and obtain the different energy consumption inputs of each device at the next moment.

[0120] Preferably, if Figure 2 As shown in the figure, a two-layer intelligent agent swarming control strategy is adopted to adjust the energy efficiency of enterprises, including: with the goal of maximizing the energy efficiency output indicators of high-energy-consuming enterprises, a second intelligent agent is set to actively track the changes in different energy inputs, and dynamically update the coordinated inputs of different energy sources based on the changes; a first intelligent agent is set to determine the energy consumption inputs of different devices and different energy sources at the next moment based on the total energy consumption per unit time of a certain type of energy provided by the second intelligent agent and the device interaction constraints associated with the graph, so as to form a strategy for increasing or decreasing the input of each device and each energy at the next moment.

[0121] Preferably, the specific scheme for adjusting the energy efficiency of an enterprise by adopting the double-layer intelligent agent swarming control strategy is:

[0122] Build device interaction constraints based on the acquired enterprise-related data.

[0123] Preferably, the device interaction constraints are constructed in the following specific manner:

[0124] In the first agent's swarming control strategy, the equipment is used as an agent, and an undirected graph is used to represent the interaction between agents:

[0125] G(t)=(I,E(t))

[0126] Wherein, G(t) represents the interaction constraint graph between production equipment at time t, where t is the control time; I represents the set of equipment nodes, I = {1, 2, …, N}, and E(t) represents the set of edges between nodes at time t. If there is an interaction constraint between node i and node j at time t, there is a connected edge between the two nodes, and its value is (a, b), where i and j represent the i-th and j-th equipment nodes respectively, i∈[1, N], j∈[1, N], and N represents the total number of equipment; a represents the delay time for the downstream node to adjust the energy consumption after the signal of the upstream production node is sent to the downstream node, and its value is obtained according to the statistics of the production completion time of the upstream node; b represents the ratio of the year-on-year increase or decrease of the upstream and downstream nodes when adjusting the energy consumption, and its value is obtained according to the statistical data of various energy consumption statistics of the upstream and downstream nodes.

[0127] It should be noted that, based on the device interaction constraints associated with the graph, energy consumption control rules are formulated to achieve consistent regulation of energy consumption between associated devices, thereby avoiding the problem of inconsistent energy consumption regulation leading to excess output value on one side and too low output value on the other side of the upstream and downstream devices.

[0128] Construct the energy consumption change control equation of the two-layer intelligent agent, set nodes and expected goals.

[0129] Preferably, the energy consumption change, nodes and expected goals of the two-layer agent are specifically:

[0130] Energy consumption change control equation:

[0131] The total number of the first agent and the second agent is Each agent is regarded as a node; its energy consumption changes conform to the following control equation:

[0132]

[0133] Among them, u α represents the control input item of the αth agent, as the unit energy consumption acceleration; v α represents the unit energy consumption of the αth following agent as the energy consumption rate; qα Represents the cumulative energy consumption of the αth following agent

[0134] node:

[0135] The first intelligent agent node is set as the device node, and the input is the energy consumption per unit time of each device with the same energy source; the second intelligent agent node is set as the node of all devices with different energy sources, and the input is the total energy consumption per unit time of all devices with different energy sources;

[0136] Expected goals:

[0137] The expected goal of the second intelligent agent is to maximize the comprehensive energy efficiency output index; the expected goal of the first intelligent agent is the total energy consumption per unit time of all devices of a certain energy source determined by the second intelligent agent.

[0138] Based on the acquired energy constraint data, a control input form of the second intelligent agent is formulated, and the second intelligent agent determines the accumulated energy consumption of different energy sources at the next moment according to the control input form.

[0139] Preferably, the second agent control input form is specifically:

[0140]

[0141] In the formula, They represent the cumulative energy consumption of the kth energy source at the next moment and this moment respectively. The cumulative amount of each energy source at the initial moment is a set random value; represents the unit energy consumption of the kth energy at the next moment and this moment, respectively, b k represents the energy input controller of the second agent, ε is the set energy efficiency evaluation threshold, which is 10 -5 ; Indicates the increase or decrease in the energy input of the second agent at the next moment; k represents the index allocation correction factor of the k-th energy source, K′ is the total number of energy categories, x k , τ k They represent the energy cost of the k-th type of energy, comprehensive energy utilization rate and energy quality satisfaction of the enterprise, represents the weight of the contributing factors to the above three indicators, and The energy distribution ratio represents the comprehensive contribution value of the kth energy minus the average contribution value of all energy sources;

[0142] The adjusted energy amount of the second intelligent body must be within the constraints. The value of must be no less than 0 and cannot exceed the output power of the kth energy source at a certain moment. The value of must be between the maximum and minimum values ​​of the total ramp rate of all devices of the k-th energy source at a time. If any value exceeds its range, its nearest boundary constraint value is taken as its value. The energy consumption control range constraint condition is expressed as:

[0143]

[0144] Among them, M1(k) represents the k-th energy constraint range, represents the upper limit of the output power of the k-th energy at a time, M2(k) represents the output power constraint range of the k-th energy at a time, and They are respectively the minimum and maximum values ​​of the total ramp rate of all devices of the k-th energy within a moment.

[0145] It should be noted that one moment is one hour; the weights of the contributing factors of the three indicators are The indicators are set according to the emphasis placed by different companies on these three indicators.

[0146] Combined with the constructed device interaction constraints, the control input form of the second agent is formulated, and the second agent determines the energy consumption input of different energy sources and different devices at the next moment based on the control input form.

[0147] It should be noted that any first agent can track a designated second agent, that is, the total energy input of the first agent will converge to the total energy input allocated and obtained by the second agent.

[0148] Preferably, the first agent control input form is specifically:

[0149]

[0150] Where Δf i g is the energy consumption control rule function f i g The derivative of (v) represents the energy consumption adjustment per unit time between intelligent agents under the constraints of graph-related interactions. The superscript g indicates that the function is an energy consumption control rule function. The device node obtains the quantitative data of the unit energy consumption and cumulative energy consumption of other device nodes associated with it based on the information interaction between undirected graphs. If there is an interaction constraint between two devices at time t, then their energy consumption control rule f i g (v) as follows:

[0151]

[0152] in, They represent the unit energy consumption of the jth device node at time t+1 and time t respectively.

[0153] They represent the unit energy consumption of the i-th device node at time t+1 and time t respectively; They represent the optimal unit energy consumption of equipment j and equipment i respectively, which are the optimal unit energy consumption set by the factory parameters of the equipment or the optimal unit energy consumption obtained from daily operation statistics; It represents the optimal unit energy consumption ratio of device j and device i with interaction constraints; It represents the unit energy consumption ratio of device j and device i at time t with interaction constraints;

[0154] In the formula, f γ i It is the target feedback item, which is used to track the expected target and calculate the difference between the expected target and the total accumulated energy consumption of each device node. The superscript γ indicates that the function is the target feedback item function. The energy consumption increase or decrease of each device is determined according to the difference and the weighted proportion of the energy consumption of each device. The specific change control equation is:

[0155]

[0156] in, represents the cumulative energy consumption of the k-th second agent, whose value is the expected goal of the first agent, and k represents the k-th type of energy; represents the unit energy consumption of the i-th first agent, is the sum of the unit energy consumption of M first agents, where M is the number of first agents in the kth energy category; the energy consumption of the first agent following the second agent, and the absolute value of the difference is within the set threshold ε′, which is 0.1% of the energy consumption of the second agent, indicating that the first agent has completed tracking the energy consumption change of the second agent, otherwise, continue to use the above change control equation to increase or decrease the unit energy consumption of each first agent in proportion to complete the change tracking;

[0157] The equipment needs to meet the energy consumption constraint. The adjusted amount of the first agent's energy must be greater than or equal to 0 and less than the maximum power limit of its input energy. The formula is:

[0158] 0≤v k,i ≤v k,max,i

[0159] Among them, v k,i represents the unit energy consumption of the kth type of energy of the ith equipment; v k,max,i Represents the maximum power limit of the energy type of the ith device.

[0160] Based on the different energy consumption inputs of each device at the next moment, adjust the energy consumption inputs of each device at the current moment, input the adjusted data into the energy efficiency evaluation model of high-energy-consuming units, and obtain the input relaxation value, output relaxation value and various output indicators;

[0161] If the output relaxation value decreases, the energy consumption at the next moment will be adjusted based on the different energy consumption inputs of each device at the next moment; if the output relaxation value does not decrease, the two-layer intelligent agent swarm control adjustment strategy will be updated, and when it is determined that the enterprise's energy efficiency unit time decision unit is in a weakly effective state at the current moment, the two-layer intelligent agent swarm control strategy will be used to adjust the enterprise's energy efficiency.

[0162] Preferably, a specific scheme for updating the double-layer agent swarming control adjustment strategy is: updating the random value set for the cumulative amount of each energy source at the initial moment when the second agent tracks the enterprise energy.

[0163] Specifically, to compare and illustrate the effectiveness of the method described in this patent, the proposed two-layer intelligent swarming control energy efficiency optimization scheme is simulated and verified in the Matlab / Simulink platform based on the 8-node equipment interaction constraint constructed by Steel Company A, and its historical data from January 2023 to January 2024 are randomly divided into 2 parts, sample set 1 is used as the sample set for evaluation before energy efficiency optimization of the enterprise, and sample set 2 is used as the sample set for evaluation of energy efficiency of the enterprise after energy efficiency regulation and optimization. Input various energy data of the sample set into the constructed energy efficiency evaluation model of high-energy-consuming energy-consuming units to obtain the mean value of the enterprise energy efficiency optimization evaluation before each typical scenario, as shown in Table 2:

[0164] Table 2 Average evaluation of enterprise energy efficiency before optimization in each typical scenario

[0165]

[0166] As shown in Table 2, it is the average value of energy efficiency evaluation for each typical scenario before the enterprise energy efficiency optimization. Among them, the enterprise energy saving cost is the evaluation index data after the enterprise energy efficiency optimization, so it is 0 at this time; the comprehensive energy utilization rate represents the complementary effect between various energy sources, so the lower the value, the better; the energy quality satisfaction is The value represents the ratio of energy consumption output to input, and the larger the value, the better; the higher the efficiency value of the timing decision unit, the better the comprehensive evaluation value of the enterprise's energy efficiency. Its maximum ideal value is 1, indicating that all links of the enterprise are in an effective state. At this time, the input relaxation value and the output relaxation value are both 0; the lower the input relaxation value and the output relaxation value, the better.

[0167] Taking scenario 1 as an example, in view of the greater random variation of solar energy and photovoltaic output during the energy tracking process, in order to demonstrate the energy tracking performance of the method described in this patent, the photovoltaic energy supply ratio and the wind turbine energy supply ratio are selected on the horizontal and vertical axes respectively, such as Figure 3The blue dots shown in the figure are the state of PV energy supply accounting for 3.5% and wind turbine energy supply accounting for 3% at a certain moment in scenario 1. Before the enterprise energy efficiency optimization, the energy consumption growth rate of the equipment new energy (including PV and wind turbine output) represented by the black nodes is as follows: Figure 3 As shown by the arrows at each node, the direction of the energy consumption growth rate makes it difficult to effectively track the changing state of new energy.

[0168] This patent adopts a double-layer intelligent agent swarming control method to extract the new energy consumption input data of each device in sample set 2 that meets the photovoltaic energy supply ratio of 3.5% and the wind turbine energy supply ratio of 3%, such as Figure 4 As shown, by adopting the patented method, the energy supply growth rate of each device can effectively track the time-varying energy supply.

[0169] By adopting the double-layer intelligent agent swarming control energy efficiency optimization to carry out energy consumption control optimization simulation test on steel enterprise A, various energy data of sample set 2 are input into the constructed energy efficiency evaluation model of high energy-consuming energy-consuming units to obtain the average evaluation value of enterprise energy efficiency optimization in each typical scenario, as shown in Table 3:

[0170] Table 3 Average evaluation of energy efficiency after optimization in each typical scenario

[0171]

[0172]

[0173] In order to compare and illustrate the effectiveness of the patent method, the energy-saving cost rate, comprehensive energy utilization rate, energy quality satisfaction, and timing decision unit efficiency values ​​of Table 2 and Table 3 are processed by difference; the input relaxation value and output relaxation value are processed by difference, and then divided by the data acquisition percentage before optimization, such as Figure 5 As shown:

[0174] like Figure 5 As shown, the energy-saving cost of the enterprise is reduced by 6.32% in the 8th scenario and by 20% in the 1st scenario; the comprehensive energy utilization rate is increased by 29.42% to 42.52%, indicating that this method can effectively improve the complementarity between different energy sources; the energy quality satisfaction is improved by 8.68% to 14.15%, thereby effectively improving the energy input and energy output ratio, indicating that the phenomenon of wind and solar power abandonment is significantly reduced. As a whole, it is shown that when the energy supply changes, the method described in this patent can track the energy changes and coordinate the energy supply of each equipment, thereby increasing the consumption of new energy and improving the degree of complementarity between different energy sources, thereby effectively reducing the production cost of A Steel Company. This patent adjusts the input relaxation value and output relaxation value of each energy of A Steel Company based on the double-layer intelligent body swarming control, and optimizes the energy consumption distribution of each equipment, thereby improving the efficiency value of the timing decision unit. Figure 5It can be seen that the input slack value is reduced by more than 64.78% to 75.39%, the output slack value is reduced by more than 32.53% to 62.87%, and the efficiency value of the timing decision unit is improved by 11.01% to 20.85, indicating that the patented method can improve the efficiency of the timing decision unit, reduce the energy input redundancy rate and the expected output shortage rate, and can flexibly adjust and expand energy according to the specific energy consumption characteristics and evaluation index requirements, and is explainable and operational, which can intuitively reflect the energy efficiency level and improvement space of the enterprise.

[0175] Embodiment 2 of the present invention provides an energy efficiency improvement system based on double-layer intelligent agent swarming control, using the energy efficiency improvement method based on double-layer intelligent agent swarming control described in Embodiment 1, including:

[0176] Sample collection module: used to obtain time-series energy consumption samples of high-energy-consuming enterprises, wherein the time-series energy consumption samples of high-energy-consuming enterprises include energy consumption characteristic data, enterprise-related data, and energy constraint data;

[0177] Energy efficiency output index construction module: used to construct energy efficiency output indicators based on the energy consumption characteristic data, wherein the energy efficiency output indicators include the enterprise energy saving cost ratio, comprehensive energy utilization rate and energy quality satisfaction;

[0178] Index output module: used to extract the different energy consumption data of the enterprise every hour based on the energy consumption characteristic data as the input index, build an energy efficiency evaluation model for high-energy-consuming energy-consuming units based on data envelopment analysis, and output the efficiency value, input relaxation value, output relaxation value and energy efficiency output index of the enterprise's time series decision unit;

[0179] Enterprise energy efficiency status judgment module: used to judge the enterprise energy efficiency status according to the input relaxation value and the output relaxation value, set the energy efficiency evaluation threshold, if the input relaxation value and the output relaxation value are both less than or equal to the energy efficiency evaluation threshold, then it is determined that the enterprise energy efficiency unit time decision unit is valid at the current moment; if the input relaxation value or the output relaxation value is greater than the set energy efficiency evaluation threshold, then it is determined that the enterprise energy efficiency unit time decision unit is in a weakly valid state at the current moment;

[0180] Enterprise energy efficiency adjustment module: if it is determined that the enterprise energy efficiency unit time decision unit is in a weak effective state at the current moment, it uses a two-layer intelligent agent swarming control strategy to adjust the enterprise energy efficiency based on the enterprise association data and energy constraint data, and obtains the different energy consumption inputs of each device at the next moment;

[0181] Enterprise energy efficiency adjustment progress judgment module: Based on the different energy consumption inputs of each device at the next moment, the energy consumption inputs of each device at the current moment are adjusted, and the adjusted data are input into the energy efficiency evaluation model of high-energy-consuming units to obtain input relaxation values, output relaxation values ​​and various energy efficiency output indicators; if the output relaxation value decreases, the energy consumption at the next moment will be adjusted based on the different energy consumption inputs of each device at the next moment; if the output relaxation value does not decrease, the two-layer intelligent agent swarm control adjustment strategy is updated, and the two-layer intelligent agent swarm control strategy based on enterprise-related data and energy constraint data is returned to adjust the enterprise energy efficiency.

[0182] The beneficial effect of the present invention is that, compared with the prior art,

[0183] (1) The present invention adopts graph association to set the interaction constraints between devices, and quantifies the interaction change constraints of energy flow between process-related devices by process delay and year-on-year rate increase or decrease. It fully considers the spatiotemporal characteristics and coupling characteristics of multiple energy flows such as cold, heat, electricity, and steam in the energy supply system of high-energy-consuming enterprises under different energy consumption boundaries, as well as the energy flow correlation and constraints between key equipment, ensuring that the energy supply control of various types of equipment is within a stable and reasonable range;

[0184] (2) The present invention adopts data envelopment analysis that meets multiple input-multiple output indicators to construct an energy efficiency evaluation model for high-energy-consuming energy-consuming units, takes time as the energy efficiency evaluation unit, selects the energy cost of the enterprise, the comprehensive energy utilization rate and the energy quality satisfaction of the enterprise for comprehensive evaluation of the energy efficiency of the enterprise, fully considers the preference of the multi-dimensional indicators of the enterprise subject, and can effectively improve the comprehensive energy efficiency of the time unit operation of high-energy-consuming enterprises by reducing the input relaxation value and the output relaxation value;

[0185] (3) The present invention adopts a two-layer intelligent agent swarm control to formulate an energy consumption control plan. The upper layer (i.e., the second intelligent agent) is responsible for determining the optimal coordinated input of various energy sources, and the lower layer (i.e., the first intelligent agent) determines a certain energy consumption input of each device based on a certain energy input determined by the upper layer. This solves the problem of complex multi-energy supply correlation and high difficulty in overall energy consumption control in high-energy-consuming enterprises. Through two-layer optimization scheduling, energy balance and optimal utilization are achieved.

[0186] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for improving energy efficiency based on double-layer intelligent agent swarming control, characterized in that: include: Obtaining time-series energy usage samples of high-energy-consuming enterprises, wherein the time-series energy usage samples of high-energy-consuming enterprises include energy usage characteristic data, enterprise-related data, and energy constraint data; Based on the energy consumption characteristic data, construct energy efficiency output indicators, wherein the energy efficiency output indicators include the enterprise energy saving cost ratio, comprehensive energy utilization rate and energy quality satisfaction; Based on the energy consumption characteristic data, extract the different energy consumption data of the enterprise every hour as the input index, build an energy efficiency evaluation model for high-energy-consuming energy-consuming units based on data envelopment analysis, and output the efficiency value, input relaxation value, output relaxation value and energy efficiency output index of the enterprise's time series decision unit; The energy efficiency status of the enterprise is judged according to the input slack value and the output slack value, and the energy efficiency evaluation threshold is set. If the input slack value and the output slack value are both less than or equal to the energy efficiency evaluation threshold, the energy efficiency unit time decision unit of the enterprise at the current moment is determined to be valid; If the input slack value or the output slack value is greater than the set energy efficiency evaluation threshold, it is determined that the energy efficiency unit time decision unit of the enterprise is in a weakly effective state at the current moment; When it is determined that the energy efficiency unit time decision unit of the enterprise is in a weak effective state at the current moment, a two-layer intelligent agent swarming control strategy is used to adjust the enterprise energy efficiency based on the enterprise association data and energy constraint data to obtain the different energy consumption inputs of each device at the next moment; Based on the different energy consumption inputs of each device at the next moment, adjust the energy consumption inputs of each device at the current moment, input the adjusted data into the energy efficiency evaluation model of high-energy-consuming units, and obtain the input relaxation value, output relaxation value and various energy efficiency output indicators; If the output slack value decreases, the energy consumption at the next moment will be adjusted based on the different energy consumption inputs of each device at the next moment; if the output slack value does not decrease, the two-layer intelligent agent swarm control adjustment strategy will be updated, and the two-layer intelligent agent swarm control strategy will be used to adjust the enterprise energy efficiency based on the enterprise-related data and energy constraint data.

2. The energy efficiency improvement method based on double-layer agent swarming control according to claim 1 is characterized by: The energy consumption characteristic data includes the energy consumption cost, energy consumption and energy power of the enterprise equipment electricity, gas, heat, cold and new energy; The enterprise-related data in the time-series energy consumption sample of the high-energy-consuming enterprise includes enterprise equipment physical topology data, enterprise equipment production-related data, and energy consumption-related data; The energy constraint data in the time-series energy consumption samples of the high-energy-consuming enterprises include the enterprise energy constraint range, the energy output power constraint range at a moment, and the minimum and maximum values ​​of the total ramp rate of different energy supplies for different equipment at a moment.

3. The energy efficiency improvement method based on double-layer agent swarming control according to claim 2 is characterized by: The enterprise energy-saving cost ratio in the energy efficiency output index is used to characterize the change in the cost ratio of electricity and gas charges at all times before and after the energy consumption regulation of the enterprise within the set time period and the electricity and gas charges at the corresponding time before the adjustment; The comprehensive energy utilization rate in the energy efficiency output index represents the degree of complementarity between different energy sources of the enterprise within a set time period; The energy quality satisfaction in the energy efficiency output index ensures the energy consumption efficiency of the enterprise.

4. The energy efficiency improvement method based on double-layer agent swarming control according to claim 3 is characterized by: The specific calculation formula for the enterprise energy saving cost ratio in the energy efficiency output index is: In the formula, C represents the energy-saving cost ratio of the enterprise; Respectively represent the energy consumption of the enterprise before and after the adjustment of the t s The change in electricity and gas charges at the data sampling time, among which, They represent the electricity and gas expenditure costs of the enterprise before the energy adjustment. They represent the electricity and gas expenditure costs of the enterprise after the energy consumption is adjusted; e represents electricity, g represents gas, and T represents a regulation time period, that is, the time period from the weak effective state before the adjustment to the end of the regulation. The T time is divided into several sampling moments, t s Indicates the data sampling time; the smaller C is, the more energy the enterprise saves and the lower the operating cost.

5. The energy efficiency improvement method based on double-layer agent swarming control according to claim 3 is characterized by: The calculation formula for the comprehensive energy utilization rate in the energy efficiency output index is specifically: Where Rc represents the comprehensive energy utilization rate; r t′ Indicates any wind, solar, hydropower, thermal power, gas from t s Time to t s The complementary coefficient at the +1 moment, the smaller the complementary coefficient, the more the fluctuation of energy consumption is offset, and the stronger the complementarity between energy sources; They represent the cumulative energy consumption fluctuation rates of wind power, photovoltaic power, hydropower, thermal power, and gas in the time window t′ respectively; Respectively represent Energy in t s +1 moment and t s Energy consumption at any time, For the Energy in [t s ,t s +1] Maximum energy consumption during the period, 6. The energy efficiency improvement method based on double-layer agent swarming control according to claim 3 is characterized by: The specific calculation formula for energy quality satisfaction in the energy efficiency output index is: Where η ex Energy consumption for enterprises Efficiency; e , gas , h , c , rn They are the energy quality coefficient weights of electricity, gas, heat, cold, and new energy, which are equal to the proportion of each energy supply cost in the historical operation of the enterprise. The sum of the energy quality coefficient weights is 1; Respectively expressed in t s The electrical load power, heating load power and cooling load power of the nth device at the moment represent the effective energy value of the device; Respectively expressed in t s The electricity, natural gas, thermal energy and new energy power input by the nth device at the moment, including solar energy input and wind energy input Same as equipment input new energy N represents the total number of devices.

7. The energy efficiency improvement method based on double-layer agent swarming control according to claim 1 is characterized by: Constructing an energy efficiency evaluation model for high energy consumption units based on data envelopment analysis DEA (λ,θ) specifically includes: Where x lκ represents the lth input indicator data of the κth decision-making unit, where κ = 1, 2, ..., K, K represents the total number of energy efficiency time decision-making units of the enterprise, l = 1, 2, ... L, L represents the total number of input indicators; y rk represents the energy efficiency output index data of the rth decision-making unit of the κth decision-making unit, where r = 1, 2, …, R, R represents the total number of output indicators; θ represents the efficiency value of the time decision-making unit; S - , S + Respectively represent the input relaxation index value and output relaxation index value; ε is 10 -5 ; l Represents the contribution weight of the lth input indicator to energy efficiency evaluation.

8. The energy efficiency improvement method based on double-layer agent swarming control according to claim 1 is characterized by: The double-layer intelligent agent swarming control strategy is used to adjust the energy efficiency of enterprises, including: with the goal of maximizing the energy efficiency output index of high-energy-consuming enterprises, setting a second intelligent agent responsible for actively tracking the changes in different energy inputs, and continuously and dynamically updating the coordinated inputs between different energy sources based on the changes; setting a first intelligent agent to determine the energy consumption inputs of different devices and different energy sources at the next moment based on the total energy consumption per unit time of a certain type of energy provided by the second intelligent agent, combined with the device interaction constraints associated with the graph, and form a strategy for increasing or decreasing the input of each device and each energy at the next moment.

9. The energy efficiency improvement method based on double-layer agent swarming control according to any one of claims 1 to 8, characterized in that: The specific scheme of adjusting the energy efficiency of enterprises by using the double-layer intelligent agent swarming control strategy based on enterprise association data and energy constraint data is as follows: Build equipment interaction constraints based on the acquired enterprise-related data; Construct the energy consumption change control equation of the two-layer intelligent agent, set nodes and expected goals; Based on the acquired energy constraint data, a control input form of the second intelligent agent is formulated, and the second intelligent agent determines the accumulated energy consumption of different energy sources at the next moment according to the control input form; In combination with the constructed device interaction constraints, a control input form of the first intelligent agent is formulated, and the first intelligent agent determines the energy consumption input of different energy sources and different devices at the next moment based on the control input form.

10. The energy efficiency improvement method based on double-layer agent swarming control according to claim 9, characterized in that: The specific method of constructing the device interaction constraint is as follows: In the first agent's swarming control strategy, the equipment is used as an agent, and an undirected graph is used to represent the interaction between agents: G(t)=(I,E(t)) Wherein, G(t) represents the interaction constraint graph between production equipment at time t, where t is the control time; I represents the set of equipment nodes, I = {1, 2, …, N}, and E(t) represents the set of edges between nodes at time t. If there is an interaction constraint between node i and node j at time t, there is a connected edge between the two nodes, and its value is (a, b), where i and j represent the i-th and j-th equipment nodes respectively, i∈[1, N], j∈[1, N], and N represents the total number of equipment; a represents the delay time for the downstream node to adjust the energy consumption after the signal of the upstream node is sent to the downstream node, and its value is obtained according to the statistics of the production completion time of the upstream node; b represents the ratio of the year-on-year increase or decrease of the upstream and downstream nodes when adjusting the energy consumption, and its value is obtained according to the statistical data of various energy consumption statistics of the upstream and downstream nodes.

11. The energy efficiency improvement method based on double-layer agent swarming control according to claim 10, characterized in that: The energy consumption changes, nodes and expected goals of the two-layer agent are specifically as follows: Energy consumption change control equation: The total number of the first agent and the second agent is Treat each agent as a node; The energy consumption changes all conform to the following control equation: Among them, u α represents the control input item of the αth agent, as the unit energy consumption acceleration; v α represents the unit energy consumption of the αth following agent as the energy consumption rate; q α Represents the cumulative energy consumption of the αth following agent node: The first intelligent agent node is set as the device node, and the input is the energy consumption per unit time of each device with the same energy source; the second intelligent agent node is set as the node of all devices with different energy sources, and the input is the total energy consumption per unit time of all devices with different energy sources; Expected goals: The expected goal of the second intelligent agent is to maximize the comprehensive energy efficiency output index; the expected goal of the first intelligent agent is the total energy consumption per unit time of all devices of a certain energy source determined by the second intelligent agent.

12. The energy efficiency improvement method based on double-layer agent swarming control according to claim 11, characterized in that: The second agent control input form is specifically: In the formula, They represent the cumulative energy consumption of the kth energy source at the next moment and this moment respectively. The cumulative amount of each energy source at the initial moment is a set random value; represents the unit energy consumption of the kth energy at the next moment and this moment, respectively, b k represents the energy input controller of the second agent, ε is the set energy efficiency evaluation threshold, which is 10 -5 ; Indicates the increase or decrease in the energy input of the second agent at the next moment; λ k represents the index allocation correction factor of the k-th energy source, K′ is the total number of energy categories, χ k , τ k They represent the energy cost of the k-th type of energy, comprehensive energy utilization rate and energy quality satisfaction of the enterprise, represents the weight of the contributing factors to the above three indicators, and The energy distribution ratio represents the comprehensive contribution value of the kth energy minus the average contribution value of all energy sources; The adjusted energy amount of the second intelligent body must be within the constraints. The value of must be no less than 0 and cannot exceed the output power of the kth energy source at a certain moment. The value of must be between the maximum and minimum values ​​of the total ramp rate of all devices of the k-th energy source within a moment. If any value exceeds its range, the nearest boundary constraint value is taken as its value.

13. The energy efficiency improvement method based on double-layer agent swarming control according to claim 12, characterized in that: The first agent control input form is specifically: Where Δf i g is the energy consumption control rule function f i g The derivative of (v) represents the energy consumption adjustment per unit time between agents under the graph-related interaction constraints. The superscript g indicates that the function is an energy consumption control rule function. If there is an interaction constraint between two devices at time t, then their energy consumption control rule f i g (v) as follows: in, They represent the unit energy consumption of the jth device node at time t+1 and time t respectively. They represent the unit energy consumption of the i-th device node at time t+1 and time t respectively; They represent the optimal unit energy consumption of equipment j and equipment i respectively, which are the optimal unit energy consumption set by the factory parameters of the equipment or the optimal unit energy consumption obtained from daily operation statistics; It represents the optimal unit energy consumption ratio of device j and device i with interaction constraints; It represents the unit energy consumption ratio of device j and device i at time t with interaction constraints; In the formula, f i γ It is the target feedback item, which is used to track the expected target and calculate the difference between the expected target and the total accumulated energy consumption of each device node. The superscript γ indicates that the function is the target feedback item function. The energy consumption increase or decrease of each device is determined according to the difference and the weighted proportion of the energy consumption of each device. The specific change control equation is: in, represents the cumulative energy consumption of the k-th second agent, whose value is the expected goal of the first agent, and k represents the k-th type of energy; represents the unit energy consumption of the i-th first agent, is the sum of the unit energy consumption of M first agents, where M is the number of first agents in the kth energy category; the energy consumption of the first agent following the second agent, and the absolute value of the difference is within the set threshold ε′, which is 0.1% of the energy consumption of the second agent, indicating that the first agent has completed tracking the energy consumption change of the second agent, otherwise, the above change control equation is continued to increase or decrease the unit energy consumption of each first agent in proportion to complete the change tracking; The equipment must meet the energy consumption constraint, and the adjusted amount of the first agent's energy must be greater than or equal to 0 and less than the maximum power limit of its input energy.

14. The energy efficiency improvement method based on double-layer agent swarming control according to claim 12, characterized in that: The specific scheme for updating the double-layer agent swarming control adjustment strategy is: updating the random value set for the cumulative amount of each energy source at the initial moment when the second agent tracks the enterprise energy.

15. An energy efficiency improvement system based on double-layer intelligent agent swarming control using the method according to any one of claims 1 to 14, characterized in that: include: Sample collection module: used to obtain time-series energy consumption samples of high-energy-consuming enterprises, wherein the time-series energy consumption samples of high-energy-consuming enterprises include energy consumption characteristic data, enterprise-related data, and energy constraint data; Energy efficiency output index construction module: used to construct energy efficiency output indicators based on the energy consumption characteristic data, wherein the energy efficiency output indicators include the enterprise energy saving cost ratio, comprehensive energy utilization rate and energy quality satisfaction; Index output module: used to extract the different energy consumption data of the enterprise every hour based on the energy consumption characteristic data as the input index, build an energy efficiency evaluation model for high-energy-consuming energy-consuming units based on data envelopment analysis, and output the efficiency value, input relaxation value, output relaxation value and energy efficiency output index of the enterprise's time series decision unit; Enterprise energy efficiency status judgment module: used to judge the enterprise energy efficiency status according to the input relaxation value and the output relaxation value, set the energy efficiency evaluation threshold, and if the input relaxation value and the output relaxation value are both less than or equal to the energy efficiency evaluation threshold, it is determined that the enterprise energy efficiency unit time decision unit is valid at the current moment; If the input slack value or the output slack value is greater than the set energy efficiency evaluation threshold, it is determined that the energy efficiency unit time decision unit of the enterprise is in a weakly effective state at the current moment; Enterprise energy efficiency adjustment module: if it is determined that the enterprise energy efficiency unit time decision unit is in a weak effective state at the current moment, it uses a two-layer intelligent agent swarming control strategy to adjust the enterprise energy efficiency based on the enterprise association data and energy constraint data, and obtains the different energy consumption inputs of each device at the next moment; Enterprise energy efficiency adjustment progress judgment module: Based on the different energy consumption inputs of each device at the next moment, the energy consumption inputs of each device at the current moment are adjusted, and the adjusted data are input into the energy efficiency evaluation model of high-energy-consuming units to obtain input relaxation values, output relaxation values ​​and various energy efficiency output indicators; if the output relaxation value decreases, the energy consumption at the next moment will be adjusted based on the different energy consumption inputs of each device at the next moment; if the output relaxation value does not decrease, the two-layer intelligent agent swarm control adjustment strategy is updated, and the two-layer intelligent agent swarm control strategy based on enterprise-related data and energy constraint data is returned to adjust the enterprise energy efficiency.

Images