A Source-Load Intelligent Matching Method for a Regional Integrated Energy System

Abstract

The present invention discloses a source-load intelligent matching method for a regional integrated energy system. According to the power curve of the load in the regional integrated energy system, the K-means algorithm is used to cluster the load; then the source-load matching algorithm is used. First, the problem of single source-load matching in the regional integrated energy system (RIES) is analyzed, and an optimization control model for the clustered load of the regional integrated energy system is constructed. Then, through the relationship between the time series of the source-load curves, the two-way restrictive attribute of source-load cointegration is utilized to further match the source-load curves, making their coupling degree higher. Different from traditional unidirectional scheduling, the cointegration model realizes the safe interactive operation of the source and load and the comprehensive and efficient utilization of resources by establishing a two-way feedback mechanism between the source and load.

Description

Technical Field

[0001] The present invention relates to the technical field of energy supply, and more specifically to a method for energy matching coupling of a regional integrated energy system. Background Art

[0002] At present, the growth rate of industrial parks' contribution to the national economy has exceeded 30%, and their energy consumption accounts for about 69% of the total energy consumption in the society, of which carbon emissions account for as high as 31% of the total emissions. Therefore, industrial parks have become the main destination for carbon neutrality.

[0003] At present, there is an unreasonable energy consumption structure in industrial regions, among which fossil energy accounts for 84% of the total regional energy consumption, and a large amount of unconventional energy has not been actively utilized. Therefore, the regional production industry effectively integrates distributed resources such as photovoltaics, wind power, and industrial waste heat and waste gas, promotes the adjustment of regional industrial structure from the perspective of multi-energy coupling, creates a new structure of comprehensive energy and low-carbon energy supply with a further increase in the proportion of renewable energy, and constructs a regional integrated energy system (RIES) with complementary energy coupling, realizes the green transformation of regional energy production from both the source and load sides, promotes the proportion of clean and low-carbon energy, and lays the foundation for building a zero-carbon park.

[0004] Then, due to the uncertainty of renewable energy itself and the time-varying characteristics of production load, both supply and demand sides show a strong trend of random changes, which leads to the difficulty in matching the energy supply on the regional production side with the real-time demand on the consumption side. The volatility of renewable energy does not match the energy consumption characteristics of production load, resulting in the abandonment rate of wind and solar power in industrial production as high as 30%, which limits the energy utilization rate and overall benefit of renewable energy in the region. On the other hand, although demand-side management incorporates large-scale production loads in the region as adjustable resources into regional energy scheduling and operation, due to the excessive dispersion of production load space, various load operation parameters have obvious peak-valley characteristics, resulting in an average load rate of only 43% for power supply equipment, and less than 5% of the time when the load rate is above 95%. The problem of low energy utilization has increased the operation and maintenance costs of various energy supply and consumption systems, resulting in a huge waste of social funds. At the same time, it is difficult for the control center to accurately perceive and precisely control the load conditions, resulting in the existing unified control mode no longer being applicable, and the quality of product mining is difficult to guarantee, which limits the speed of regional low-carbon transformation and development.

[0005] Therefore, how to achieve active coordination between sources and loads in a regional integrated energy system and perform refined control of supply and demand matching and load is an issue that technical personnel in this field urgently need to solve. Summary of the invention

[0006] In view of this, the present invention provides a source-load intelligent matching method for a regional integrated energy system, which is a method for matching the load and energy supply of an industrial park system. Aiming at overcoming the waste problem caused by the mismatch between the source and the load in the regional integrated energy system, a source-load power curve matching optimization method based on an artificial intelligence algorithm is adopted, so as to carry out refined control of the supply-demand and load, avoid energy waste, and reduce carbon emissions.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions:

[0008] A source-load intelligent matching method for a regional integrated energy system includes the following steps:

[0009] Step 1: Collect historical load data in the regional integrated energy system, and use the K-means algorithm to perform load clustering on the historical load data to obtain several groups of load clusters;

[0010] An adaptive K-means algorithm that can automatically select the number of clustering centers is used to group the loads in the regional integrated energy system (RIES) according to the power curve shape, find the power consumption units with similar power consumption rules (i.e., each power-consuming part), and divide the units with similar properties into one group for subsequent lean adjustment control; the basic idea of K-means is to assign the power of the load curve at each measurement moment to randomly selected K load centers, repeat this behavior and update until the grouping result does not change and the curve error sum is the smallest;

[0011] Step 2: Divide the loads in each group of load clusters into adjustable loads and fixed loads according to production characteristics; and preset a power adjustable range for the time-varying loads in the adjustable loads;

[0012] Divide the cluster loads P in the region Ln into adjustable loads P according to production characteristics I and fixed loads P T ; Adjustable loads are loads that can operate with peak shifting and intermittently, such as pumping units, cast iron furnaces, etc.; Fixed loads are loads that need to run continuously before the work task is completed, such as copper melting shaft furnaces; Time-varying loads are loads that change with time. For the time-varying loads in the adjustable clusters after clustering, the power adjustable range of each cluster is specified, and the expression is as follows:

[0013]

[0014] Among them, P t I is the cluster load power at time t, They are the maximum and minimum values of the cluster load power at the next control moment respectively; i is the working state of the load at time t, where i = 0 indicates that the load is in the shutdown state and i = 1 indicates that the load is in the startup state; in the traditional working mode, all the production loads in each cluster are started and work without interruption, and the total power is at the maximum value. Therefore, there is a large downward adjustment space for the cluster load power.

[0015] S I,min ≤S t I ≤S I,max

[0016] Among them, S I,max and S I,min are the upper and lower limits of the regulation capacity of the cluster load respectively.

[0017] For the load with regulation capacity, the threshold of its shutdown time needs to be considered to ensure that the production process in the region will not be affected during the load regulation operation. Therefore, the cluster load shutdown time T I The constraint conditions are shown in the following formula:

[0018] 0 ≤ T I ≤ T I,max

[0019] Among them, T I,max is the maximum stoppable working time of the adjustable load P I .

[0020] The load power on the bus of the time-varying load of a certain cluster at any moment can be expressed by the following formula:

[0021]

[0022] Step 3: Build a matching model, match the source power curve and the load power curve according to the objective function of the matching model. If the matching result meets the constraint conditions of the matching model, no adjustment is made and the matching continues; otherwise, go to Step 4 for power adjustment.

[0023] The matching is to perform power matching considering the load on the load side, and match the load power curve corresponding to each adjustable load with the source power curve in turn.

[0024] The closer the source power curve and the load power curve are, the higher the matching degree; if the difference between the source power curve and the load power curve at the same moment is less than 0.4, they are matched.

[0025] Historical load data includes the source power curve and the load power curve; the source power curve represents the power supply law, and the load power curve represents the electricity consumption law of the loads within the region; after clustering and grouping the load power curve using the K-means algorithm, the firefly algorithm utilizes its own algorithmic advantages to match the source power curve and the load power curve, obtaining the gap between the power supply law and the electricity consumption law, thereby achieving the description of the high-dimensional non-linear relationship between the source and load within the region; while overcoming the dynamic changes of non-stationary sequences, seeking the supply-demand constraint relationship within the region to establish a functional relationship that reflects the long-term two-way constraints of the RIES source and load variables, namely the matching model, including the objective function and the constraint conditions;

[0026] The greater the degree of coincidence between the source output curve (source power curve) and the load power curve of the regional integrated energy system, the more it can ensure the safe and stable operation of the system. Therefore, the source-load matching degree θ of the regional integrated energy system is defined. The larger θ is, the better the tracking and smoothing effect of the RIES source output curve on the load power curve; the source-load tracking coefficient δ p Mainly controls the fitting degree of the power output of the main control power source to the load; the source-load fluctuation coefficient δ S Mainly controls the smoothness of the remaining load power curve. The combination of the two can deeply depict the fluctuation difference between the source output curve and the load power curve, and construct the objective function as follows:

[0027] Objective function:

[0028]

[0029]

[0030]

[0031] C S,t =P L,t -E W,t

[0032]

[0033] α1 + α2 = 1

[0034] In the formula: P L,t is the electricity load at time t, and P L,av is the average load consuming the electricity of the RIES system; E W,t is the total output of the regional integrated energy system; E W,av is the average value of the total output of the regional integrated energy system; C S,t is the remaining load at time t; C S is the average remaining load; T is the total number of time periods; α1 and α2 are the source-load tracking coefficient δ p and the source-load fluctuation coefficient δ SThe weight coefficient, the larger its value indicates that the corresponding index is more important; generally, the initial setting is α1:α2 = 1:1;

[0035] The constraint conditions are as follows:

[0036] Power balance:

[0037] In the formula: is the total system output at time t; N V,t and N D,t are the photovoltaic output and the power supply side output at time t respectively; Δt is the time interval;

[0038] Channel limit:

[0039] In the formula: N L,max is the maximum transmission capacity of the transmission channel.

[0040] Output limit:

[0041] In the formula: and are the minimum and maximum values of the photovoltaic output at time t respectively; and are the maximum and minimum values of the power supply side output at time t respectively.

[0042] Ramp constraint:

[0043] In the formula: η i is the ramp capacity of the i-th unit.

[0044] Real-time power balance: P B,t ≤N V,t +N D,t ,

[0045] In the formula: P B,t is the load at the load end of the regional integrated energy system at time t;

[0046] Step 4: According to the constraint conditions of the matching model, use the firefly algorithm to adjust the load power curve according to the preset power adjustable range, adjust the power deviation of the time-varying load within each load cluster to the preset power adjustable range, establish a time-varying load intermittent control model, and optimize the working state of the adjustable loads participating in the scheduling within each cluster;

[0047] Collect the actual working condition data based on load production and establish the RIES energy supply - demand balance model; collect the power curves of time - varying loads, and combine with the intermittent working strategy of the RIES energy supply - demand balance model to establish a time - varying load intermittent control model, control the working states of the time - varying loads within each group of load clusters, adjust the start - up and shut - down times of adjustable loads, and adjust the power deviation of the time - varying loads within each group of load clusters to the preset range.

[0048] The working states include the on - state and the off - state; after satisfying the regional constraint relationships, according to the actual working conditions of load production, establish a time - varying load intermittent control model without changing the working process. The idea of time - varying load intermittent control is to analyze the operating states of each independent time - varying load in the cluster of time - varying loads. If the demand changes, the time - varying working states participating in the scheduling within each cluster can be on or off. The purpose is to adjust the working power curves of the source and load to be as consistent as possible, so as to achieve the optimal control of the load working state.

[0049] In actual operation, the total energy output is not only affected by past values but also disturbed by variables such as load, weather, and real - time electricity price. Therefore, in this invention, the RIES source - load power is used as the main variable for supply - demand matching, and influencing factors such as weather and real - time electricity price are used as disturbance terms to establish the following RIES energy supply - demand balance model:

[0050] P Σ,t =c + dP Σ,t-1 +eP L,t +fP L,t-1 +μ t

[0051] In the formula, c, d, e, and f are all constants, P L,t , P L,t-1 are the load values in the RIES at time t and its previous moment respectively;

[0052] For the past value of energy output, use the backward - shift operator B for equivalent conversion. Let the constant term be α and the coefficient composed of the backward - shift operator be β, that is, the expression is represented as:

[0053] P Σ,t =α + βP L,t +μ t

[0054] In the actual RIES, there are multiple groups of output loads. Therefore, expand P Σ,t =α + βP L,t +μ t to the multi - output load sequence P Ln =(P Ln,1 , P Ln,2 ,..., P Ln,t)When all the cluster time-varying loads within the coverage area are considered, the RIES energy supply-demand balance model under the participation of multi-cluster loads is as follows:

[0055]

[0056] Using the constraint conditions of the matching model to control the working state of the time-varying loads within each load cluster, and adjusting the start-up and shut-down times of the adjustable loads, the output limit is expressed as:

[0057]

[0058] In the formula: and are the minimum and maximum values of the PV output at time t, respectively; and are the maximum and minimum values of the power output on the power supply side at time t, respectively.

[0059] The ramping constraint is:

[0060]

[0061] In the formula: η i is the ramping ability of the i-th unit;

[0062] The output limit and ramping constraint also affect the operating states of the individual time-varying loads in the cluster time-varying load. When the power output on the power supply side increases and the ramping ability enhances, the operating states of the loads within the cluster need to be changed accordingly, that is, they need to be turned on. The time-varying working states participating in the scheduling can be judged;

[0063] In the time-varying load intermittent control model of the regional integrated energy system, the real-time power balance: P B,t ≤N V,t +N D,t , The constraint is crucial. When the load P B,t at the load end becomes smaller, for the power output on the power supply side, it is necessary to control the real-time power balance, that is, for each individual time-varying load in the cluster time-varying load, it needs to be reduced accordingly, then the operating state needs to be turned off accordingly, and the time-varying load intermittent control module is a simple judgment of the operating states of the individual time-varying loads in the cluster load of the regional integrated energy system.

[0064] As can be seen from the above technical solutions, compared with the prior art, the present invention discloses a method for intelligent matching of sources and loads in a regional integrated energy system. According to the power curve of the load in the regional integrated energy system, the K-means algorithm is used to cluster the load; then the source-load matching algorithm is used. First, the problem of single source-load matching in the regional integrated energy system (RIES) is analyzed, and an optimization control model for the clustered load of the regional integrated energy system is constructed. Then, through the relationship between the time series of the source-load curves, using the two-way constraint attribute of source-load cointegration, the source-load curves are further matched to make their coupling degree higher. Different from the traditional one-way scheduling, a long-term stable relationship exists between the source and the load, which is in a cointegration state. The cointegration model establishes a two-way feedback mechanism between the source and the load, adjusts the consistency of the electricity consumption rules on the load side, and adjusts the electricity load of the adjustable load to match the power supply and consumption, thereby reducing energy efficiency and realizing the safe interaction operation of the source and the load and the comprehensive and efficient utilization of resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only the embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to the provided drawings.

[0066] Figure 1 The drawings are the flowcharts of the method for intelligent matching of sources and loads in the regional integrated energy system provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0067] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some of the embodiments of the present invention, rather than all of them. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0068] An embodiment of the present invention discloses a method for intelligent matching of sources and loads in a regional integrated energy system, and the flowchart is as Figure 1 shown.

[0069] S1: Cluster and group the loads in the regional integrated energy system (RIES);

[0070] For the historical load data in the RIES, an adaptive k-means algorithm that can automatically select the number of cluster centers is used to group the loads in the RIES according to the shape of the power curve;

[0071] In the K-means algorithm, K represents clustering the data into K clusters, and means represents the mean value of the data in each cluster as the center of the cluster, also known as the centroid. The K-means clustering attempts to group similar objects into the same cluster and dissimilar objects into different clusters. Here, a calculation method for measuring the similarity of data is required. The K-means algorithm is a typical distance-based clustering algorithm, using distance as the evaluation index of similarity, and by default, the Euclidean distance is used as the similarity measure, that is, the closer the distance between two objects, the greater their similarity;

[0072] The idea of the K-means algorithm is to cluster with K points in space as the centers and classify the objects close to them. Through an iterative method, the values of the clustering centers (centroids) are updated successively until the best clustering result is obtained;

[0073] The specific process of the RIES load clustering grouping is as follows:

[0074] S11: First, select the center points of K load categories;

[0075] S12: For the power samples of any load curve at each measurement moment, calculate its distance to the load centers of each category, and classify the sample into the category where the load center with the shortest distance is located;

[0076] S13: After clustering, recalculate the position of the center point of each cluster;

[0077] S14: Repeat steps S12 - S13 for iteration until the positions of the K load center points remain unchanged, or a certain number of iterations are reached, the RIES load grouping result does not change and the curve error sum is the smallest, then the iteration ends, otherwise continue the iteration;

[0078] S2: Divide the clustered loads in the area;

[0079] The clustered loads in the area are divided into adjustable loads and fixed loads according to production characteristics; adjustable loads are loads that can operate with peak shaving and intermittently, such as pumping units, cupola furnaces, etc.; fixed loads are loads that need to run continuously before the work task is completed, such as copper melting shaft furnaces; and for the time-varying loads in the clusters of adjustable loads after clustering, the adjustable range of the power of each cluster is specified;

[0080] S3: Area internal source-load intelligent matching algorithm;

[0081] Build a matching model for matching;

[0082] S4: Time-varying load optimal control;

[0083] Intelligent algorithms have the advantages of parallel efficiency, strong adaptability, and no need for special information in solving multi-objective problems. They are divided into heuristic algorithms and bionic algorithms. Since any two fireflies in the Firefly Algorithm can be compared with each other, they have strong global search performance, so the success rate of optimization is higher when dealing with high-dimensional nonlinear relationships between sources and loads in the region.

[0084] The core idea of the firefly algorithm is that each firefly has relative fluorescence brightness and attractiveness. The relative brightness can indicate the quality of the target value, and the attractiveness is used to determine the update distance of the firefly position. These two indicators are continuously iterated to finally obtain the optimization result. The algorithm is described from a mathematical perspective as follows:

[0085] S41: The formula for relative fluorescence brightness I is:

[0086]

[0087] Where: I0 is the autofluorescence brightness when R=0 at the light source, that is, the maximum fluorescence brightness, and the autofluorescence brightness is proportional to the objective function value; γ is the light absorption coefficient, which is used to reflect the characteristics of fluorescence weakening with increasing distance and medium absorption. Theoretically, γ∈[0,∞), but in practical applications, γ∈[0.01,∞); R ij is the distance between fireflies i and j, that is

[0088]

[0089] x i 、x j are the positions of fireflies i and j respectively; x ik represents the spatial coordinates of firefly i, that is, the k-dimensional coordinate value of firefly i; x jk represents the k-dimensional coordinate value of firefly j; k, d are the dimensions of the coordinate system;

[0090] S42: The absorbance w of fireflies is expressed as:

[0091]

[0092] Where: w0 is the maximum attraction, that is, the attraction at the light source (R = 0), and in most cases w0 = 1;

[0093] S43: Location update;

[0094] x i =x i +w(x j -x i )+με i (3)

[0095] Where: xi , x j are the positions of fireflies i and j respectively; μ is the step size factor, which is a constant on [0, 1]; ε i represents a random vector subject to Gaussian distribution or uniform distribution; με i is the update position perturbation term;

[0096] The firefly algorithm is used to make the value of the source-load matching degree θ of the regional integrated energy system larger, that is, to make the source-end output curve of the regional integrated energy system more consistent with the load power curve; the source-load tracking coefficient δ p and the source-load fluctuation coefficient δ S can be reflected by the relative fluorescence brightness and attractiveness of fireflies, and finally the value of the source-load matching degree θ is adjusted by updating the positions of fireflies to achieve the expected result;

[0097] The firefly algorithm (Firefly Algorithm) is used to describe the high-dimensional non-linear relationship between the source and load in the region, seek the supply-demand constraint relationship in the region while overcoming the dynamic changes of non-stationary sequences, so as to establish a functional relationship reflecting the long-term two-way constraints of the source-load variables of the RIES;

[0098] Aiming at the actual working conditions based on load production, a time-varying load intermittent control model is established without changing the working process. Through the analysis of the operating states of each independent time-varying load and combined with the intermittent working strategy, the working states of the time-varying loads participating in the scheduling in each cluster are optimized and controlled.

[0099] In this specification, each embodiment is described in a progressive manner. The key point of each embodiment is to illustrate the differences from other embodiments. The same or similar parts among the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the description of the method part.

[0100] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A source-load intelligent matching method for a regional integrated energy system, characterized in that, Including the following steps: Step 1: Collect historical load data in the regional integrated energy system, and use the K-means algorithm to perform load clustering on the historical load data to obtain several groups of load clusters; the historical load data includes a source power curve and a load power curve; Step 2: Divide the loads within each group of load clusters into adjustable loads and fixed loads according to production characteristics; and preset a power adjustable range for the time-varying loads among the adjustable loads; Step 3: Construct a matching model, and match the source power curve and the load power curve according to the objective function of the matching model. If the matching result meets the constraint conditions of the matching model, no adjustment is made and the matching continues; otherwise, go to Step 4 for power adjustment; Use the firefly algorithm to match the source power curve and the load power curve, obtain the gap between the power supply law and the power consumption law, and realize the description of the high-dimensional non-linear relationship between the source and load in the region; while overcoming the dynamic changes of non-stationary sequences, seek the supply-demand constraint relationship in the region, establish a functional relationship reflecting the long-term two-way constraints of the RIES source-load variables, and obtain a matching model. The matching model includes an objective function and constraint conditions; Source-load tracking coefficient δ p and source-load fluctuation coefficient δ S Combined with an in-depth characterization of the fluctuation differences between the source power curve and the load power curve, a target function is constructed, expressed as: θ represents the source-load matching degree of the regional integrated energy system, and α1 and α2 are the source-load tracking coefficient δ p and the source-load fluctuation coefficient δ S respectively; the constraint conditions include power balance, channel limit, output limit and ramp constraint; Step 4: According to the constraint conditions of the matching model, use the firefly algorithm to adjust the load power curve according to the preset power adjustable range, adjust the power deviation of the time-varying loads within each group of load clusters to within the preset power adjustable range, establish a time-varying load intermittent control model, and optimize the control of the working states of the adjustable loads participating in the scheduling within each cluster; Collect the actual working condition data based on load production and establish a RIES energy supply-demand balance model; collect the power curve of the time-varying load, and combine the intermittent working strategy of the RIES energy supply-demand balance model to establish a time-varying load intermittent control model, control the working states of the time-varying loads within each group of load clusters, adjust the start-stop times of the adjustable loads, and adjust the power deviation of the time-varying loads within each group of load clusters to within the preset range.

2. The source-load intelligent matching method for a regional integrated energy system according to claim 1, wherein In Step 1, group the loads corresponding to the power curves in the historical load data, assign the power of the power curve of the load at each measurement moment to K load centers, and obtain K groups of load groupings.

Images