Compressor cluster distributed collaborative scheduling method based on contract energy management

Abstract

The application discloses a compressor cluster distributed collaborative scheduling method based on contract energy management, and particularly relates to the technical field of industrial energy saving. First, the compressor cluster is divided into distributed nodes to construct a distributed collaborative scheduling network, then each intelligent node of the compressor collects the running parameters of itself in real time through the built-in sensor to construct a compressor energy consumption model and obtain the predicted energy consumption of the intelligent node; next, a distributed optimization algorithm is adopted to obtain the load distribution result of each intelligent node and the total energy consumption of the region; finally, the global energy saving accounting result of the compressor is obtained through the income accounting rules in the contract energy management system, and is summarized and fed back; through the distributed collaborative scheduling, the real-time running state and local load demand of each compressor can be known, the optimized operation of the compressor is realized, and the energy utilization efficiency of the whole compressor cluster is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial energy saving, in particular to a compressor cluster distributed collaborative scheduling method based on contract energy management. BACKGROUND

[0002] In industrial production, compressors, as important power equipment, are widely used in refrigeration, gas delivery and other fields. With the continuous expansion of industrial scale, the application of compressor clusters is becoming more and more common. Contract energy management is a new market-based energy saving mechanism, which provides energy saving services for customers and shares energy saving benefits through energy management contracts.

[0003] The existing compressor scheduling method mainly focuses on single optimization algorithm (such as PSO, fuzzy PID) or device balanced operation, balances the service life of compressors by adjusting the start-up sequence, or virtual power plant distributed collaboration, optimizes resource scheduling through two-way communication and intelligent algorithm, or adopts centralized control, that is, a central controller is used to uniformly schedule and manage all compressors.

[0004] However, there are many problems in the operation and management of the compressor cluster at present. On the one hand, the traditional centralized scheduling method is slow in response when facing complex and variable production demands, and it is difficult to accurately match the output of each compressor with the actual energy demand in real time, resulting in serious energy waste. For example, in some production scenarios, due to the inability of centralized scheduling to timely perceive local energy fluctuations, some compressors may be in an overrunning or inefficient running state. On the other hand, the existing scheduling method often ignores the individual differences of compressor devices. Different models and different service life of compressors have significant differences in energy consumption characteristics and load adaptability, but the existing scheduling strategy fails to fully utilize these characteristics for optimal scheduling. Against this background, how to realize efficient distributed collaborative scheduling of compressor clusters within the framework of contract energy management, improve energy utilization efficiency and reduce energy cost has become a technical problem to be solved. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a compressor cluster distributed collaborative scheduling method based on contract energy management to solve the problems raised in the background.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical scheme: a compressor cluster distributed collaborative scheduling method based on contract energy management, comprising:

[0007] S1: divide the compressor group into distributed nodes, each distributed node is configured with a corresponding local controller, at the same time set a global coordinator, and initialize the parameter setting of the compressor independent intelligent node in each distributed node, and construct a distributed collaborative scheduling network;

[0008] S2: Each compressor intelligent node collects its own operation parameters in real time through the built-in sensor and stores them in the local database to obtain a set of operation parameters of the compressor cooperative scheduling;

[0009] S3: A compressor energy consumption model is constructed based on the set of operation parameters of the historical compressor cooperative scheduling, the set of operation parameters of the real-time compressor cooperative scheduling is input into the compressor energy consumption model, and the predicted energy consumption of the intelligent node is obtained;

[0010] S4: A distributed optimization algorithm is used to process the predicted energy consumption of the intelligent node received by the local controller to obtain the load distribution result of each intelligent node in the region and the total energy consumption of the region;

[0011] S5: The global coordinator optimizes the energy saving benefit based on the total energy consumption of the region through the benefit accounting rules in the contract energy management system to obtain the global energy saving accounting result of the compressor;

[0012] S6: Based on the benefit accounting period in the contract energy management system, the global energy saving accounting result of the compressor is summarized to obtain the period cumulative energy saving benefit and judge whether it meets the expectation, and the abnormal result is fed back to the management terminal for human-computer interaction.

[0013] The technical effects and advantages of the present application are as follows:

[0014] 1. The present application can realize the optimal operation of the compressor according to the real-time operation state and local load demand of each compressor through distributed cooperative scheduling, avoid the overloading operation of part of the compressors and the light load or idle operation of part of the compressors, realize the reasonable distribution of the compressor cluster load, and thus improve the energy utilization efficiency of the entire compressor cluster;

[0015] 2. The present application adopts the distributed cooperative scheduling strategy, avoids the single point failure risk of the centralized control mode, each compressor node has the ability of cooperative scheduling, even if part of the nodes fail, it will not affect the normal operation of the entire system, improves the reliability and stability of the system; at the same time, the local controller and the global coordinator ensure that the entire system can quickly adjust the scheduling strategy when facing complex and variable production demands, maintain stable operation, and improve the adaptability and reliability of the system;

[0016] 3. The present application can accurately calculate the energy saving benefit according to the contract agreed energy saving benchmark and energy saving amount calculation method, realize real-time and accurate accounting of energy saving benefit, reduce energy consumption and energy cost through optimizing the operation of the compressor under the contract energy management mode, and bring higher energy saving benefit to customers and energy saving service companies. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 Fig. 1 is a schematic diagram of the overall process of the present application.

[0018] Figure 2 Fig. 2 is a schematic diagram of the process of the present application. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0020] Please refer to Figure 1 As shown in Fig. 1, the present application provides a compressor cluster distributed collaborative scheduling system based on contract energy management, which comprises a distributed collaborative scheduling network construction module, a compressor intelligent node data acquisition module, a compressor intelligent node energy consumption analysis module, a regional distributed collaborative scheduling module, a compressor global optimization control module, and an energy saving benefit accounting and feedback module.

[0021] The distributed collaborative scheduling network construction module is connected with the other modules, the compressor intelligent node data acquisition module is connected with the compressor intelligent node energy consumption analysis module, the regional distributed collaborative scheduling module is connected with the compressor intelligent node energy consumption analysis module and the compressor global optimization control module respectively, and the energy saving benefit accounting and feedback module is connected with the compressor global optimization control module and the regional distributed collaborative scheduling module respectively.

[0022] The distributed collaborative scheduling network construction module is used for dividing the target compressor group into distributed nodes, each distributed node is configured with a corresponding local controller, a global coordinator is set at the same time, and the initialization parameter setting is performed on the independent intelligent nodes of the compressors in each distributed node to construct the distributed collaborative scheduling network.

[0023] The compressor intelligent node data acquisition module: each compressor intelligent node acquires the running parameter set of itself in real time through the built-in sensor, stores in the local database at the same time, obtains the running parameters of the collaborative scheduling of the compressors, and transmits to the compressor intelligent node energy consumption analysis module.

[0024] The compressor intelligent node energy consumption analysis module: based on the running parameter set of the historical collaborative scheduling of the compressors, a compressor energy consumption model is constructed, the running parameter set of the real-time collaborative scheduling of the compressors is input into the compressor energy consumption model, the predicted energy consumption of the intelligent node is obtained, and is transmitted to the regional distributed collaborative scheduling module.

[0025] Regional distributed collaborative scheduling module: a distributed optimization algorithm is used to process the predicted energy consumption of the intelligent nodes received by the local controller, to obtain the load distribution results of each intelligent node in the region and the total energy consumption of the region, and to transmit the total energy consumption of the region to the compressor global optimization control module;

[0026] Compressor global optimization control module: the global coordinator optimizes the energy saving benefit based on the total energy consumption of the region through the benefit accounting rules in the energy management system, obtains the global energy saving accounting result of the compressor, and transmits it to the energy saving benefit accounting and feedback module;

[0027] Energy saving benefit accounting and feedback module: based on the benefit accounting period in the energy management system, the global energy saving accounting result of the compressor is summarized to obtain the period cumulative energy saving benefit and to judge whether it meets the expectation, and the abnormal result is fed back to the management terminal for human-computer interaction.

[0028] Please refer to Figure 2 Based on the compressor cluster distributed collaborative scheduling method in the energy management, as shown in the figure, including: S1: the compressor group is divided into distributed nodes, each distributed node is configured with a corresponding local controller, a global coordinator is set at the same time, and the initialization parameter setting of the independent intelligent node of the compressor in each distributed node is performed, and a distributed collaborative scheduling network is constructed; S2: each compressor intelligent node collects the running parameters of itself in real time through the built-in sensor, and stores them in the local database, to obtain the running parameter set of the compressor collaborative scheduling; S3: based on the historical running parameter set of the compressor collaborative scheduling, a compressor energy consumption model is constructed, the real-time running parameter set of the compressor collaborative scheduling is input into the compressor energy consumption model, and the predicted energy consumption of the intelligent node is obtained; S4: a distributed optimization algorithm is used to process the predicted energy consumption of the intelligent nodes received by the local controller, to obtain the load distribution results of each intelligent node in the region and the total energy consumption of the region; S5: the global coordinator optimizes the energy saving benefit based on the total energy consumption of the region through the benefit accounting rules in the energy management system, obtains the global energy saving accounting result of the compressor; S6: based on the benefit accounting period in the energy management system, the global energy saving accounting result of the compressor is summarized to obtain the period cumulative energy saving benefit and to judge whether it meets the expectation, and the abnormal result is fed back to the management terminal for human-computer interaction.

[0029] S1: the compressor group is divided into distributed nodes, each distributed node is configured with a corresponding local controller, a global coordinator is set at the same time, and the initialization parameter setting of the independent intelligent node of the compressor in each distributed node is performed, and a distributed collaborative scheduling network is constructed, including the following steps:

[0030] S1.1: Divide the target compressor group into n distributed nodes, each containing at least one compressor and configured with a corresponding local controller, n>1, and set up a global coordinator;

[0031] It is particularly pointed out in this embodiment that each distributed node can contain one or more compressors; the local controller has data acquisition, processing and communication functions, can obtain the running parameters of the compressor in real time, such as pressure, temperature, flow, power, etc., and can interact with adjacent nodes; the global coordinator is used to coordinate the scheduling decisions between the distributed nodes.

[0032] S1.2: Each compressor in each distributed node is regarded as an independent intelligent node and initialized with parameters, including the number, rated power, rated flow, current running state and current load of the compressor, the current running state including start, running and stop; at the same time, the distributed collaborative scheduling parameters are set, including the scheduling period (for example, 8 minutes), the communication interval (for example, 1 minute) and the collaborative optimization goal of minimizing total energy consumption and maximizing energy saving benefit;

[0033] S1.3: Configure the communication network between the distributed nodes through the wireless communication module (such as zigBee, Wi-Fi, etc.), and build a distributed collaborative scheduling network;

[0034] S2: Each compressor intelligent node collects its own running parameters in real time through the built-in sensor and stores them in the local database, obtaining the running parameter set of the compressor collaborative scheduling, which includes input power, exhaust flow, exhaust pressure, running temperature and motor current;

[0035] It is particularly pointed out in this embodiment that the sensor includes but is not limited to temperature sensor, pressure sensor, flow sensor, etc.

[0036] S3: Based on the historical compressor collaborative scheduling running parameter set, a compressor energy consumption model is built, and the real-time compressor collaborative scheduling running parameter set is input into the compressor energy consumption model to obtain the predicted energy consumption of the intelligent node, including the following steps:

[0037] S3.1: Build a compressor energy consumption model:

[0038] S3.1.1: Through N sets of historical compressor collaborative scheduling running parameter set D of the i-th intelligent node and the corresponding n sets of energy consumption set E, the running parameter set D includes the input power P, exhaust flow Q, exhaust pressure F, motor current I and running temperature T of the intelligent node; then the N sets of i-th intelligent node running parameter set Dy and energy consumption set Ey are obtained by preprocessing the running parameter set D and energy consumption set E;

[0039] It needs to be specifically pointed out that the preprocessing operation is the prior art, including denoising processing, outlier processing and data standardization processing; denoising can use filtering algorithms such as median filtering, mean filtering, etc.; outlier processing can use statistical methods such as the 3σ principle; and the collected historical data is standardized by the Z-score standardization method.

[0040] S3.1.2: First, the least squares support vector machine (LS-SVM) algorithm is used to learn the mapping relationship of the compressor energy consumption model Ey LS-SVM of the set of running parameters Dy to the set of energy consumptions Ey m , to obtain the model parameters Lagrange multiplier α k , kernel function parameter γ and bias term b; , N is the number of samples, that is, N groups, Dy m and Dy k represent the kth and mth sets of intelligent node running parameters, respectively, k m , k∈n, m∈n, K(Dy k , Dy m ) is a radial basis function, K(Dy k , Dy m )=exp(-γ||Dy 2 -Dy k ||), ||Dy m -Dy LS-SVM || is the Euclidean distance; then the energy consumption model Ey act,m is iteratively trained by minimizing the error ΔE, , Ey m and Ey th are the actual energy consumption and the predicted energy consumption of the mth set of intelligent node running parameters, respectively, when ΔE is less than the corresponding threshold ΔE LS-SVM (threshold values can be obtained by statistical analysis of historical data), the training is stopped, and the trained compressor energy consumption model Ey LS-SVM is output, , L is the number of support vectors, L≤N;

[0041] It needs to be specifically pointed out that LS-SVM is the prior art, which replaces the inequality constraint in the traditional SVM with an equality constraint, converts the quadratic programming problem into a linear equation system for solving, and significantly reduces the computational complexity. The core idea is to map the original data to a high-dimensional feature space through a kernel function, and use linear SVM for classification or regression in this space.

[0042] S3.2: Input the real-time running parameter set Dy into the trained compressor energy consumption model Ey i , to obtain the predicted energy consumption Ey(t);

[0043] S4: Adopting a distributed optimization algorithm, processing the predicted energy consumption of the intelligent nodes received by the local controller, obtaining the load distribution results of each intelligent node in the region and the total energy consumption of the region, including the following steps:

[0044] S4.1: Adopting a distributed optimization algorithm (for example, ADMM algorithm), processing the predicted energy consumption of the intelligent nodes received by the local controller, taking the minimization of the total energy consumption of the region as the objective function, and taking the load distribution results of each intelligent node as the constraint condition, to construct the objective function formula of the distributed optimization algorithm: , E j,tot is the total energy consumption of the jth distributed node, N j is the number of intelligent nodes in the jth distributed node, Ey i (t) is the predicted energy consumption of the ith intelligent node, L i,j (t) is the load allocated by the jth distributed node to the ith intelligent node (for example, L i,j (t)=50m 3 / min, indicating that the load allocated by the jth distributed node to the ith intelligent node at time t is "compress and output 50 cubic meters of gas per minute"), the load is the gas compression work to be completed (unit: m 3 / min), L req,j (t) is the total gas demand of the jth distributed node (for example, the total gas demand of a factory at a certain time period is 200m 3 / min); , C i (t) is the load carrying capacity of the ith intelligent node at time t, , Q i,des is the design flow of the ith intelligent node, which is the maximum exhaust capacity under the rated working condition marked when the equipment is shipped, P i,ra is the rated power of the ith intelligent node, which is the maximum input power marked on the motor nameplate, P i,los is the additional loss power of the ith intelligent node, P i,los =k f ×f i (t)+k T ×(T i (t)-T i,base ), f i (t) is the motor frequency of the ith intelligent node at time t (unit: Hz), k f is the frequency loss coefficient (unit: kW / Hz, determined by the motor characteristics, for example, 0.2kW / Hz), T i (t) and T i,baseLet k be the operating temperature and reference temperature of the i-th smart node at time t (unit: °C, reference temperature is usually taken as 40 °C). T For a temperature-induced energy loss system (unit: kW / ℃, e.g., 0.5 kW / ℃, which can be obtained by fitting historical temperature-energy consumption data; the higher the temperature, the greater the energy loss), η i,mech Let η be the mechanical efficiency (range 0 to 1) of the i-th intelligent node at time t. i,mech =η i,ra ×(1-k η ×|F i (t)-F i,opt |), η i,ra F is the mechanical efficiency under rated operating conditions. i (t) and F i,opt Let k be the exhaust pressure of the i-th smart node at time t and the exhaust pressure corresponding to the optimal efficiency (unit: MPa). η Efficiency attenuation coefficient (unit: MPa) -1 For example, 0.2 MPa -1 The efficiency is obtained by fitting historical exhaust pressure-mechanical efficiency data (the further the deviation from the optimal pressure, the lower the efficiency). i,safe (t) represents the safety coefficient of the i-th smart node at time t. T i,max I is the maximum safe temperature allowed by the compressor. i (t) and I i,ra Let be the motor current and rated current of the i-th smart node at time t, respectively;

[0045] This embodiment needs to specifically explain that the distributed optimization algorithm (ADMM algorithm) is an existing technology. It is a distributed iterative algorithm used to solve separable convex optimization problems. The core idea is to decompose the target problem into multiple sub-problems. In this application, the global optimization problem of minimizing the total energy consumption of the region is decomposed into sub-problems (local optimization) of each compressor, and then coordination is achieved through "coordination variables" (such as intermediate variables of load distribution within the region).

[0046] S4.2: Verify the load L of the i-th smart node i,j (t) satisfies the constraint condition [L] i,min ,L i,max In the case of L i,min and L i,max Given the minimum and maximum loads of the i-th smart node, respectively, the load L of each compressor is iteratively optimized using the residual convergence function R. i,j (t) and the total regional energy consumption E is iteratively optimized through the energy consumption deviation coefficient η(ΔE). j,tot The residual convergence function R includes the total regional load deviation r. p Deviation r from distributed optimization algorithmd , , ||·|| is the L2 norm, ρ is the penalty parameter of the distributed optimization algorithm, and L i,j new (t) and L i,j old (t) represents the load distribution of smart nodes in adjacent iteration steps; E j,tot new (t) and E j,tot old (t) represents the total energy consumption of the distributed nodes in adjacent iteration steps; when L i,j (t)∈[L i,min ,L i,max ]、r p ≤r p th And r d ≤r d th And η(ΔE)≤η(ΔE) th When both conditions are met, r p th r d th and η(ΔE) th These are the corresponding preset thresholds (e.g., r). p th =10 -3 m 3 / min、r d th =10 -3 m 3 / min and η(ΔE) th =10 -2 This indicates that the distributed optimization algorithm has converged, and finally found the load distribution results of each smart node that meets the total gas demand and the total energy consumption of the distributed nodes. The total energy consumption of the distributed nodes is the total energy consumption of the region.

[0047] S5: The global coordinator optimizes energy-saving benefits based on the total regional energy consumption using the revenue accounting rules in the energy performance contracting management system, and obtains the global energy-saving accounting results for the compressor, including the following steps:

[0048] S5.1: Based on the revenue accounting rules in the Energy Performance Contracting (EPC) management system, the energy-saving benefits are optimized according to the total regional energy consumption to obtain the energy-saving benefit function maxB(t) for global optimization and control. maxB(t) represents the maximum energy saving benefit at time t, n is the number of distributed nodes in the region, and E j,base (t) and E j,tot(t) is the energy-saving benchmark energy consumption (agreed value in the contract energy management system) and the predicted total energy consumption of the distributed node j at time t, respectively, C en C is the energy unit price, and cr is the energy-saving benefit sharing ratio (for example, cr = 0.6, which means that the energy service company ESCO accounts for 60%);

[0049] S5.2: Iterative optimization is performed on the global optimization regulated energy-saving benefit function maxB(t), and when the global energy-saving benefit change rate coefficient η(B(t)) ≤ the corresponding threshold η(B(t)) th (derived through historical data statistical analysis), E j,tot (t) ≥ the minimum energy consumption E j,min (t) of the distributed node j,base (t) ≥ E j,tot (t) of the distributed node en C is the energy unit price, and 0 < cr < 1; Then, maxB(t) is compared with the historical optimal value, and if the difference Δ(maxB(t)) ≥ 0 and meets the expected growth range, it is determined that the energy-saving benefit function maxB(t) converges, and the compressor global energy-saving accounting result is obtained. Otherwise, return to step S5.1 for iterative optimization. , maxB new (t) and maxB old (t) are the maximum energy-saving benefits of adjacent iterative steps, respectively.

[0050] S6: Based on the benefit accounting period in the contract energy management system, the compressor global energy-saving accounting result is summarized to obtain the period cumulative energy-saving benefit and determine whether it meets the expectation. Abnormal results are fed back to the management terminal for human-computer interaction, including the following steps:

[0051] S6.1: In the benefit accounting period Tt (the period can be days, months, etc. according to the contract), the compressor global energy-saving accounting result is summarized to obtain the period cumulative energy-saving benefit maxB(Tt), , maxB(t) is the maximum energy-saving benefit at time t, M is the sampling number in the period, and J is the Jth sampling.

[0052] S6.2: Calculate the mean mu(B(Tt)) and standard deviation sigma(B(Tt)) of the cumulative energy saving benefits of the history m1 periods, if the period yield rate η(B(Tt)) ∈ [mu(B(Tt))-3sigma(B(Tt)), mu(B(Tt))+3sigma(B(Tt))], it is explained that the period cumulative energy saving benefit meets the expectation, otherwise the abnormal period cumulative energy saving benefit is fed back to the management terminal, and the calculation logic of the energy saving benefit function maxB(t) of the global optimization control, the optimization distributed optimization algorithm and the like are checked reversely.

[0053] Secondly: the drawings in the disclosed embodiments of the application only involve structures related to the disclosed embodiments of the application, other structures can refer to general design, and in the case of no conflict, the same embodiment and different embodiments of the application can be combined with each other;

[0054] Finally: the above only describes the preferred embodiments of the application, and is not used to limit the application, and any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the application should be included in the protection scope of the application.

Claims

1. A distributed collaborative scheduling method for compressor clusters based on energy performance contracting, characterized by: include: S1: Divide the compressor group into distributed nodes, each distributed node is configured with a corresponding local controller, and a global coordinator is set up. Initialize the parameters of the independent intelligent nodes of the compressors in each distributed node to build a distributed collaborative scheduling network. S2: Each compressor intelligent node collects its own operating parameters in real time through built-in sensors and stores them in a local database to obtain a set of operating parameters for compressor collaborative scheduling; S3: Construct a compressor energy consumption model based on the set of historical compressor collaborative scheduling operating parameters, input the set of real-time compressor collaborative scheduling operating parameters into the compressor energy consumption model, and obtain the predicted energy consumption of the intelligent node; The S3 implementation includes the following steps: S3.1: Constructing a compressor energy consumption model: S3.1.1: Obtain the set of operating parameters D and the corresponding set of energy consumption E of the historical compressor coordinated scheduling of the i-th intelligent node in N groups. The set of operating parameters D includes the input power P, exhaust flow rate Q, exhaust pressure F, motor current I, and operating temperature T of the intelligent node; then, obtain the set of operating parameters Dy and the set of energy consumption Ey of the i-th intelligent node in N groups by performing preprocessing operations on the set of operating parameters D and the set of energy consumption E. S3.1.2: First, the least squares support vector machine algorithm is used to learn the compressor energy consumption model Ey, which maps the set of operating parameters Dy to the energy consumption set Ey. LS-SVM The model parameters Lagrange multipliers α are obtained. m Kernel function parameter γ and bias term b; N is the number of samples, i.e., N groups, Dy k and Dy m Let K and M represent the sets of operating parameters for the k-th and m-th intelligent nodes, respectively, where k < m, k ∈ n, m ∈ n, and K(Dy k ,Dy m K(Dy) is a radial basis function. k ,Dy m )=exp(-γ||Dy k -Dy m || 2 ),||Dy k -Dy m || represents the Euclidean distance; then, the energy consumption model Ey is optimized by minimizing the error ΔE. LS-SVM Perform iterative training. Ey act,m and Ey m These are the actual energy consumption and predicted energy consumption of the m-th intelligent node's operating parameter set, respectively. When ΔE is less than the corresponding threshold ΔE... th When the time is up, stop training and output the completed compressor energy consumption model Ey. LS-SVM , L is the number of support vectors, L≤N; S3.2: Input the real-time operating parameter set Dy into the trained compressor energy consumption model Ey. LS-SVM In the process, the predicted energy consumption Ey of the i-th intelligent node is obtained. i (t); S4: Using a distributed optimization algorithm, the predicted energy consumption of the smart nodes received by the local controller is processed to obtain the load distribution results of each smart node in the region and the total energy consumption of the region. The implementation of S4 includes: S4.1: Using a distributed optimization algorithm, the predicted energy consumption received by the local controller from the intelligent nodes is processed, with the objective function being minimizing the total regional energy consumption, and the load distribution results of each intelligent node being used as constraints, to construct the objective function formula of the distributed optimization algorithm: E j,tot Let N be the total energy consumption of the j-th distributed node in the region. j Let Ey be the number of intelligent nodes within the j-th distributed node. i (t) represents the predicted energy consumption of the i-th smart node, L i,j (t) represents the load allocated by distributed node j to the i-th intelligent node, where the load is the amount of gas compression work to be completed, L. req,j (t) represents the total gas demand of distributed node j; C i (t) represents the load capacity of the i-th smart node at time t. Q i,des Let P be the design flow rate for the i-th smart node. i,ra P is the rated power of the i-th smart node. i,los Let η be the additional power loss of the i-th smart node. i,mech Let λ be the mechanical efficiency of the i-th intelligent node at time t. i,safe (t) represents the safety coefficient of the i-th intelligent node at time t; S5: The global coordinator optimizes the energy-saving benefits based on the total regional energy consumption by using the revenue accounting rules in the energy performance contracting management system, and obtains the global energy-saving accounting results for the compressor. S6: Based on the revenue calculation cycle in the energy performance contracting management system, summarize the global energy saving calculation results of the compressor, obtain the cumulative energy saving revenue of the cycle, determine whether it meets expectations, and feed back abnormal results to the management terminal for human-computer interaction.

2. The distributed collaborative scheduling method for compressor clusters based on energy performance contracting as described in claim 1, characterized in that: The implementation of S1 includes the following steps: S1.1: Divide the target compressor group into n distributed nodes. Each distributed node contains at least one compressor and is configured with a corresponding local controller. n>1. At the same time, set up a global coordinator. S1.2: Treat each compressor in each distributed node as an independent intelligent node and initialize its parameters, including the compressor number, rated power, rated flow, current operating status, and current load. The current operating status includes start, running, and stopped. At the same time, set the distributed collaborative scheduling parameters, including the scheduling cycle, communication interval, and the collaborative optimization objective of minimizing total energy consumption and maximizing energy-saving benefits. S1.3: Configure the communication network between distributed nodes through the wireless communication module to build a distributed collaborative scheduling network.

3. The distributed collaborative scheduling method for compressor clusters based on energy performance contracting as described in claim 1, characterized in that: The implementation of S4 further includes: S4.2: verifying the load L of the i-th intelligent node i,j (t) satisfies the constraint condition [L] i,min ,L i,max In the case of L i,min and L i,max Given the minimum and maximum loads of the i-th smart node, respectively, the load L of each compressor is iteratively optimized using the residual convergence function R. i,j (t) and the total regional energy consumption E is iteratively optimized through the energy consumption deviation coefficient η(ΔE). j,tot The residual convergence function R includes the total regional load deviation r. p Deviation r from distributed optimization algorithm d When L i,j (t)∈[L i,min ,L i,max ]、r p ≤r p th And r d ≤r d th And η(ΔE)≤η(ΔE) th When both conditions are met, r p th r d th and η(ΔE) th The corresponding preset thresholds indicate that the distributed optimization algorithm has converged, and finally finds the load distribution results of each smart node that meets the total gas demand and the total energy consumption of the distributed nodes. The total energy consumption of the distributed nodes is the total energy consumption of the region.

4. The distributed collaborative scheduling method for compressor clusters based on energy performance contracting as described in claim 1, characterized in that: S5 includes: S5.1: Optimizing energy-saving benefits based on the total regional energy consumption using the revenue accounting rules in the energy performance contracting management system to obtain the globally optimized energy-saving benefit function maxB(t). maxB(t) represents the maximum energy saving benefit at time t, n is the number of distributed nodes in the region, and E j,base (t) and E j,tot (t) represents the baseline energy consumption and predicted total energy consumption of distributed node j at time t, respectively. en Here, denoted as the unit price of energy, and cr represents the percentage of energy-saving benefits shared.

5. The distributed collaborative scheduling method for compressor clusters based on energy performance contracting as described in claim 1, characterized in that: The implementation in S5 also includes: S5.2: Iteratively optimizing the energy-saving benefit function maxB(t) of global optimization control, when the global energy-saving benefit change rate coefficient η(B(t)) ≤ the corresponding threshold η(B(t)). th E j,tot (t)≥Minimum energy consumption of distributed nodes E j,min (t) and E j,base (t)≥E j,tot (t), Energy unit price C en >0 and meets the contractual agreement, 0 < cr < 1; then compare maxB(t) with the historical best value. If the difference Δ(maxB(t)) ≥ 0 and meets the expected growth range, it means that the energy-saving benefit function maxB(t) has converged, and the global energy-saving calculation result of the compressor is obtained. Otherwise, return to step S5.1 to continue iterative optimization.

6. The distributed collaborative scheduling method for compressor clusters based on energy performance contracting as described in claim 1, characterized in that: The implementation of S6 includes: S6.1: Within the revenue calculation period Tt, summarizing the global energy-saving calculation results of the compressor to obtain the cumulative energy-saving revenue maxB(Tt) for the period. maxB(t) is the energy-saving benefit at time t, M is the number of samplings within the period, and J is the Jth sampling. S6.2: Calculate the mean μ(B(Tt)) and standard deviation σ(B(Tt)) of the cumulative energy saving benefits over the historical m1 cycles. If the cycle return rate η(B(Tt))∈[μ(B(Tt))-3σ(B(Tt)),μ(B(Tt))+3σ(B(Tt))], it indicates that the cumulative energy saving benefits of the cycle meet expectations. Otherwise, the cumulative energy saving benefits of the abnormal cycle are fed back to the management terminal.

Images