A municipal gas transmission and distribution integrated energy-saving system

Abstract

This invention provides a municipal gas transmission and distribution integrated energy-saving system, comprising a sensing layer, a reference model layer, a core control layer, an execution layer, and a feedback layer that interact sequentially, forming a closed-loop data link between each layer. The core objective of this invention is to address the pain points of traditional municipal gas transmission and distribution systems, which, due to their fixed parameters and manual adjustment mode, cannot adapt to fluctuations in gas composition, sudden changes in user load, and dynamic changes in pipeline impedance. These issues result in low control accuracy, large energy efficiency fluctuations, and insufficient pressure energy recovery. Ultimately, this invention aims to drive the municipal gas transmission and distribution industry towards intelligent and energy-saving upgrades.

Description

Technical Field

[0001] This invention relates to the field of municipal gas transmission and distribution and automatic control technology, and in particular to a municipal gas transmission and distribution combined energy-saving system. Background Technology

[0002] Municipal gas transmission and distribution systems are core infrastructure for urban energy supply. Their operation is characterized by significant nonlinearity, time-varying nature, and uncertainty: gas composition fluctuates with the gas source and the batch of gas being transported (methane content fluctuates between 85% and 98%); user energy load changes abruptly with time of day (significant differences between morning and evening peak hours), season (winter load is 2-3 times that of summer load), and business type (different energy consumption patterns for residential, commercial, and industrial sectors); and pipeline impedance changes dynamically with pressure, flow rate, and pipeline aging.

[0003] Traditional gas transmission and distribution systems employ a control mode of fixed parameter controllers plus manual adjustment. The core drawback is that the control parameters cannot adapt to changes in system operating conditions in real time: load prediction errors generally exceed 15%, and pipeline pressure fluctuates greatly; pipeline transmission and distribution energy consumption fluctuates by up to 30%, and pressure energy recovery efficiency is only 60%-70%; when faced with fluctuations in gas composition or sudden load changes, manual intervention is required to adjust parameters, which results in a delayed response and is prone to causing operational instability problems, making it difficult to meet the high requirements of accuracy, energy saving, and stability for modern municipal gas transmission and distribution. Summary of the Invention

[0004] This invention provides a municipal gas transmission and distribution integrated energy-saving system. Its core objective is to address the pain points of traditional municipal gas transmission and distribution systems, which rely on fixed parameters and manual adjustment. These systems are unable to adapt to fluctuations in gas composition, sudden changes in user load, and dynamic changes in pipeline impedance, resulting in low control accuracy, large energy efficiency fluctuations, and insufficient pressure energy recovery. Specific objectives include: constructing an adaptive control framework adapted to gas transmission and distribution scenarios, enabling adaptive adjustment of control parameters according to system operating conditions; improving load forecasting and pressure control accuracy; reducing pipeline transmission and distribution energy consumption and improving pressure energy recovery efficiency; enhancing system robustness to achieve stable operation without manual intervention; and ultimately promoting the intelligent and energy-saving upgrade of the municipal gas transmission and distribution industry.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] A municipal gas transmission and distribution integrated energy-saving system includes a sensing layer, a reference model layer, a core control layer, an execution layer, and a feedback layer that interact sequentially. A closed-loop data link is formed between each layer. The sensing layer collects gas composition data and environmental parameter data, which are transmitted to the reference model layer as input data. The reference model layer integrates a variational mode decomposition algorithm and an improved kernel extreme learning machine algorithm. Based on the input data, the reference model layer outputs categorized user load reference values ​​and thermal efficiency reference values. These load reference values ​​and thermal efficiency reference values ​​are transmitted to the core control layer as baseline input data. The core control layer integrates... The algorithm employs a force search algorithm and a fuzzy adaptive control algorithm. The core control layer outputs adaptive control parameters based on the reference input data and the deviation data transmitted from the feedback layer. These adaptive control parameters are transmitted to the reference model layer, the perception layer, and the execution layer. The execution layer adjusts the operating state based on the adaptive control parameters and outputs operating state data. This operating state data is transmitted to the feedback layer. The feedback layer calculates deviation data based on the operating state data and the user's actual energy consumption data. This deviation data is transmitted to the core control layer. The user's actual energy consumption data originates from the instantaneous gas consumption, cumulative gas consumption, and energy consumption period distribution data collected by smart gas meters deployed at the user's end.

[0007] In this specification, the sensing layer includes a gas component detection unit and an environmental parameter acquisition unit. The gas component detection unit detects gas components including methane, ethane, propane, n-butane, isobutane, nitrogen, and carbon dioxide. The environmental parameter acquisition unit is a built-in temperature and humidity sensor and an atmospheric pressure sensor. The environmental parameters acquired by the environmental parameter acquisition unit include temperature, humidity, and atmospheric pressure.

[0008] In this specification, the variational mode decomposition algorithm of the reference model layer decomposes the user's actual energy consumption data into multiple stationary intrinsic mode functions. The improved kernel extreme learning machine algorithm uses the intrinsic mode functions, gas composition data, and low calorific value data after environmental parameter correction as input features. The reference model layer is equipped with an error feedback mechanism. When the prediction error of the improved kernel extreme learning machine algorithm exceeds a preset threshold, the number of decomposition layers of the variational mode decomposition algorithm is adjusted and the user's actual energy consumption data is decomposed again.

[0009] In this specification, the gravity search algorithm of the core control layer takes the highest user thermal efficiency, the largest pressure energy recovery power generation, and the lowest pipeline transmission and distribution energy consumption as the multi-objective optimization direction. The optimization variables of the gravity search algorithm include the number of decomposition layers of the variational mode decomposition algorithm, the regularization coefficient of the improved kernel extreme learning machine algorithm, the pipeline impedance value, and the turbine speed value.

[0010] In this specification, the fuzzy adaptive control algorithm of the core control layer uses load deviation and thermal efficiency deviation as input variables, and load reference value correction coefficient and gravitational constant correction value of gravitational search algorithm as output variables. The fuzzy rule base of the fuzzy adaptive control algorithm contains 9 rules covering all combinations of input variables. The load deviation is the percentage difference between the load reference value and the user's actual energy consumption data, and the thermal efficiency deviation is the percentage difference between the average thermal efficiency and the preset target thermal efficiency.

[0011] In this specification, the execution layer includes a pipeline impedance adjustment unit and a pressure energy recovery and stabilization unit. The pipeline impedance adjustment unit includes an intelligent throttle valve, an adjustable damping ring, and a flexible inner wall bushing. The pipeline impedance adjustment unit adjusts the opening degree of the intelligent throttle valve, the diameter of the adjustable damping ring, and the roughness of the flexible inner wall bushing based on the pipeline impedance target value output by the core control layer.

[0012] In this specification, the pressure energy recovery and stabilization unit adopts a micro turbine generator and voltage stabilization integrated machine. The pressure energy recovery and stabilization unit adjusts the turbine speed based on the turbine speed target value output by the core control layer. The pressure energy recovery and stabilization unit adjusts the opening of the pressure stabilizing valve according to the pipeline pressure data and the user demand pressure data. The user demand pressure data comes from the rated pressure parameters of various types of user energy-consuming equipment.

[0013] In this specification, the feedback layer includes a full-link data acquisition unit and a deviation calculation unit. The data acquisition unit acquires gas composition data and environmental parameter data from the sensing layer, load reference values ​​and thermal efficiency reference values ​​from the reference model layer, and pipeline pressure data and power generation data from the execution layer. The deviation data calculated by the deviation calculation unit includes load deviation and thermal efficiency deviation.

[0014] In this specification, the core control layer is equipped with a stability verification unit. The stability verification unit uses a Lyapunov function to verify the global asymptotic stability of the system. The Lyapunov function includes load error variables, thermal efficiency error variables, and parameter estimation error variables.

[0015] In this specification, the system is equipped with an iterative optimization mechanism, which includes retraining the gravity search algorithm and fuzzy adaptive control algorithm weekly, retraining the variational mode decomposition algorithm and the improved kernel extreme learning machine algorithm monthly, and recalibrating the system stability quarterly through a stability verification unit.

[0016] In summary, the present invention has at least the following beneficial effects:

[0017] Significantly improved control accuracy: load prediction accuracy ≥98%, far exceeding the error level of traditional systems by more than 15%; pipeline pressure fluctuations are controlled within ±0.0001MPa, and the pressure stabilization accuracy meets the energy needs of various users.

[0018] The energy efficiency optimization effect is outstanding: the energy consumption of pipeline transmission and distribution is reduced by 20%-25%, and the power generation of pressure energy recovery is increased by 15%-20%; the user thermal efficiency is stable at 93%-96%, which is significantly improved compared with the standard thermal efficiency of traditional systems (90% for residential, 88% for commercial, and 92% for industrial).

[0019] Enhanced robustness and adaptability: It can adapt to fluctuations of ±10% in gas composition and sudden changes of ±30% in user load, maintaining stable system operation without human intervention, and adapting to complex and ever-changing municipal transmission and distribution conditions.

[0020] Significant economic benefits: The daily power generation of a single pressure regulating station is ≥200kWh, which can fully cover the daily power demand of the pressure regulating station, saving ≥50,000 yuan in electricity costs annually and reducing the operating costs of gas transmission and distribution. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a schematic diagram of the municipal gas transmission and distribution integrated energy-saving system involved in this invention.

[0023] Figure 2 This is a schematic diagram of the adaptive control process of the municipal gas transmission and distribution combined energy-saving system involved in this invention.

[0024] Figure 3 This is a schematic diagram of the algorithm collaborative optimization process involved in this invention.

[0025] Figure 4 This is a schematic diagram of the precise control process of execution layer pipeline impedance and pressure energy recovery involved in this invention. Detailed Implementation

[0026] In the following description, only certain exemplary embodiments are briefly described. As those skilled in the art will recognize, the described embodiments can be modified in various ways without departing from the spirit or scope of the embodiments of the invention. Therefore, the drawings and description are considered to be exemplary in nature and not restrictive.

[0027] The following disclosure provides many different implementations or examples for carrying out different structures of the embodiments of the present invention. To simplify the disclosure of the embodiments of the present invention, specific examples of components and arrangements are described below. Of course, these are merely examples and are not intended to limit the embodiments of the present invention. Furthermore, reference numerals and / or reference letters may be repeated in different examples of the embodiments of the present invention; such repetition is for simplification and clarity and does not in itself indicate a relationship between the various implementations and / or arrangements discussed.

[0028] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0029] refer to Figure 1 and Figure 2 This embodiment provides a municipal gas transmission and distribution integrated energy-saving system. The scheme designs an adaptive control energy-saving system for municipal gas transmission and distribution that integrates Variational Mode Decomposition (VMD), Improved Kernel Extreme Learning Machine (IKELM), Gravity Search Algorithm (GSA), and Fuzzy Adaptive Control Algorithm (FAC). The system adopts a five-layer closed-loop architecture: perception layer, reference model layer, core control layer, execution layer, and feedback layer. The perception layer collects disturbance data such as gas composition and environmental parameters; the reference model layer constructs reference models for different types of user loads and thermal efficiency; the core control layer achieves global optimization of control parameters and real-time deviation correction; the execution layer completes pipeline impedance regulation and pressure energy recovery; and the feedback layer ensures closed-loop data flow throughout the entire chain. Ultimately, this achieves adaptive optimal operation of the gas transmission and distribution system under nonlinear, time-varying, and uncertain operating conditions.

[0030] I. Core Positioning of the System

[0031] Municipal gas transmission and distribution systems are core infrastructure for urban energy supply. Their operation is characterized by three main features: nonlinearity, time-varying nature, and uncertainty. Gas composition fluctuates depending on the gas source and the batch being transported (e.g., methane content can range from 85% to 98%). User energy load changes abruptly with different time periods (morning peak 7:00-9:00, evening peak 18:00-20:00), seasons (winter load is 2-3 times that of summer), and business types (significant differences in energy consumption patterns between residential, commercial, and industrial sectors). Pipeline impedance dynamically changes with pressure, flow rate, and pipeline aging. Traditional gas transmission and distribution systems use a control mode with fixed parameters and manual adjustment. The controller parameters cannot adapt to real-time system state changes, resulting in low control accuracy (load prediction errors generally exceed 15%), large energy efficiency fluctuations (pipeline transmission and distribution energy consumption fluctuations reach 30%), and insufficient pressure energy recovery (recovery efficiency is only 60%-70%).

[0032] To address the aforementioned issues, this solution designs a full-link adaptive control energy-saving system based on adaptive control theory, integrating variational mode decomposition (VMD), improved kernel extreme learning machine (IKELM), gravity search algorithm (GSA), and fuzzy adaptive control algorithm (FAC). The core innovation of this solution lies in defining the entire gas transmission and distribution process as the controlled object. Through a closed-loop logic of state perception, reference modeling, global optimization, real-time correction, execution feedback, and parameter self-tuning, it achieves adaptive adjustment of control parameters in response to system disturbances (gas composition) and operating deviations (load / thermal efficiency), possessing parameter self-tuning, adaptive deviation correction, and multi-variable collaborative control capabilities.

[0033] II. System Overall Architecture

[0034] The system adopts a five-layer adaptive closed-loop architecture consisting of a perception layer, a reference model layer, a core control layer, an execution layer, and a feedback layer. Each layer forms a seamless collaborative link through data input, instruction output, and feedback correction. The core logic is as follows:

[0035] 1. The sensing layer collects disturbance data such as gas composition and environmental parameters to provide an input reference for the entire adaptive system;

[0036] 2. The reference model layer constructs a reference model for load and thermal efficiency based on disturbance data and historical operating data, and outputs the baseline target value for adaptive control;

[0037] 3. The core control layer serves as the core carrier, achieving global optimization of control parameters through GSA, and then completing adaptive correction of real-time deviations through FAC, outputting the optimal control parameters;

[0038] 4. The execution layer receives control parameters, adjusts physical parameters such as pipeline impedance and turbine speed, and completes the implementation of control commands;

[0039] 5. The feedback layer collects the running data from the execution layer, calculates the real-time deviation, and sends it back to the core control layer to achieve parameter self-tuning.

[0040] III. Detailed Design of Each Module

[0041] (a) Sensing layer: Real-time detection module for gas composition (adaptive control disturbance input source)

[0042] The core function of this module is to accurately collect the core disturbance parameters (gas composition, environmental parameters) of the gas transmission and distribution system, providing unbiased input data for subsequent adaptive control. Its detection accuracy directly determines the effectiveness of adaptive control.

[0043] 1. Input Data Collection Specifications

[0044] The input data is divided into two categories, and the collection criteria are as follows:

[0045] Gas medium samples: directly taken from the main pipeline of the city gate station, with the sampling point located 10 meters after the gate station outlet valve (to avoid component stratification caused by valve throttling), the sampling frequency is 1 time / minute, and the single sampling volume is 50 mL (to meet the detection requirements of gas chromatography-mass spectrometry).

[0046] Environmental parameters: temperature ( ),humidity( Atmospheric pressure The data is collected by a high-precision sensor built into the module. The sensor accuracy is ±0.1℃ (temperature), ±1%RH (humidity), and ±0.1kPa (atmospheric pressure). The acquisition frequency is synchronized with the gas sample (1 time / minute). The acquired data is digitally filtered (sliding window length is 5 minutes) before being output to eliminate instantaneous fluctuation interference.

[0047] 2. Core Processing Procedure

[0048] (1) Qualitative and quantitative detection of gas components

[0049] Seven core components were detected using gas chromatography-mass spectrometry (GC-MS): methane, ethane, propane, n-butane, isobutane, nitrogen, and carbon dioxide. The detection process consisted of three steps:

[0050] 1. Sample pretreatment: The gas sample is vaporized through the injection port (vaporization temperature is 100℃ to ensure complete vaporization of all components), and then carried into the chromatographic column by the carrier gas (high-purity helium, flow rate 1mL / min);

[0051] 2. Component separation: The chromatographic column used was an HP-5MS capillary column (30m in length, 0.25mm in inner diameter). The column temperature program was as follows: initial temperature 40℃ (hold for 5 min), increased to 100℃ at 5℃ / min (hold for 2 min), and increased to 200℃ at 10℃ / min (hold for 3 min). Due to differences in polarity and molecular weight, different components had different retention times in the chromatographic column (methane 1.2 min, ethane 1.8 min, propane 2.5 min, n-butane 3.2 min, isobutane 3.0 min, nitrogen 1.0 min, carbon dioxide 2.0 min), achieving complete separation.

[0052] 3. Quantitative Calculation: The separated components are introduced into a mass spectrometer detector (electron impact ion source, ionization energy 70 eV). The volume fraction of each component is calculated by matching the characteristic spectra of the ion fragment peaks with a standard spectral library. ( Up to 7, corresponding to the 7 categories of components mentioned above), the calculation formula is:

[0053] ;

[0054] In the formula, For the first Chromatographic peak areas of the components (unit: The output is directly from the mass spectrometer detector. For the first The correction factors for each component (methane 1.00, ethane 1.02, propane 1.05, n-butane 1.08, isobutane 1.07, nitrogen 0.98, carbon dioxide 1.10) were determined through standard gas calibration experiments.

[0055] (2) Calculation of the lower calorific value of gas

[0056] Step 1: Calculate the uncorrected lower heating value of the gas The formula is:

[0057] ;

[0058] In the formula, The lower heating value of the gas is the uncorrected value (unit: MJ / m³). For the first The standard lower heating values ​​of the components are as follows: methane 35.8 MJ / m³, ethane 63.8 MJ / m³, propane 91.1 MJ / m³, n-butane 118.4 MJ / m³, isobutane 117.6 MJ / m³, nitrogen 0 MJ / m³, and carbon dioxide 0 MJ / m³. For the first Volume percentage of each component (unit: %).

[0059] Calculation example: If detected (methane) (Ethane) (propane), (n-Butane) (Isobutane) (Nitrogen gas) (Carbon dioxide), then:

[0060] MJ / m³.

[0061] Step 2: Calculate the corrected lower heating value of the gas To eliminate the influence of environmental parameters on low calorific value, the formula is:

[0062] ;

[0063] In the formula, The corrected lower heating value of the gas (unit: MJ / m³). , , These are correction factors for temperature, humidity, and atmospheric pressure (calibrated in the laboratory). MJ / (m³·℃) MJ / (m³·%RH) MJ / (m³·kPa)); , , For real-time collected environmental parameters; , , This is the standard environmental benchmark value.

[0064] Calculation example: If real-time environmental parameters , , ,but:

[0065] MJ / m³.

[0066] 3. Output and cross-module integration

[0067] The module ultimately outputs two types of data: the volume percentage of the seven components. ( Up to 7); Corrected lower calorific value of gas .

[0068] The above data is synchronously transmitted to the reference model layer (as a disturbance input for the load / thermal efficiency reference model) and the core control layer (as a disturbance parameter for adaptive control) via industrial Ethernet (Modbus TCP / IP protocol, transmission delay <100ms), ensuring that subsequent modules can adapt to changes in gas characteristics in real time.

[0069] (ii) Reference Model Layer: Load Forecasting and Adaptive Reference Model Module (Adaptive Control Baseline Input Source)

[0070] refer to Figure 3 This module serves as the reference model for the adaptive control energy-saving system. Its core function is to construct load and thermal efficiency reference models for different types of users based on disturbance data from the sensing layer and historical energy consumption data from users, and to output the baseline target value for adaptive control. The module integrates variational mode decomposition (VMD) and an improved kernel extreme learning machine (IKELM), and achieves adaptive correction of the reference model through error feedback to ensure the accuracy of the baseline target value.

[0071] 1. Input data parsing

[0072] The input data contains three categories, and the preprocessing specifications are as follows:

[0073] Perturbation data: output of the perception layer ( Up to 7) The data sampling frequency is 1 time per minute, and the data is entered after time alignment (matching according to minute-level timestamps);

[0074] User energy consumption data: Sourced from smart gas meters deployed at all user terminals. User types include residential, commercial, and industrial users. The smart gas meters sample once every 15 minutes, and the collected data includes: instantaneous gas consumption. (Unit: m³ / h) Cumulative gas consumption (Unit: m³), ​​Energy Consumption Period (Unit: h); Data preprocessing includes: outlier removal (removing values ​​exceeding 3 times the standard deviation) and missing value completion (using linear interpolation).

[0075] Feedback correction data: Correction coefficient for load reference value output by the core control layer ( Representing residential users Representing business users, (Representing industrial users), used for real-time correction of reference model output.

[0076] 2. Core Algorithm Design

[0077] (1) Variational Mode Decomposition (VMD): Preprocessing of Non-stationary Load Data

[0078] The core function of VMD is to decompose non-stationary user load data into multiple stationary intrinsic mode functions (IMFs), eliminate nonlinear interference from load fluctuations, and provide high-quality input for subsequent IKELM forecasting.

[0079] ① Model building process

[0080] The core of VMD is to construct and solve variational problems, with the goal of transforming raw load data... Decomposed into A stable IMF component ( to Each component corresponds to a load fluctuation characteristic (such as intraday short-term fluctuation, intraweek medium-term fluctuation, and seasonal long-term fluctuation). The variational problem is constructed as follows:

[0081] ;

[0082] ;

[0083] In the formula, For the first One IMF component (unit: m³ / h); For the first The center frequency (in Hz) of each IMF component is used to distinguish the frequency range of different fluctuation characteristics; The number of decomposition layers (initially set to 5, and subsequently adjusted adaptively based on error feedback); Original user energy load data (unit: m³ / h); This is the Dirac function, used to characterize pulse signals; This is the convolution operator; These are the partial derivatives in the time domain; The imaginary unit; It is the square of the L2 norm, used to measure the energy of the signal.

[0084] ② Model training process

[0085] The variational problem described above is solved using the Alternating Direction Multiplier Method (ADMM), with the training dataset consisting of 72 consecutive hours of raw load data. The training steps are as follows:

[0086] 1. Initialization parameters: Set the initial number of decomposition layers. Convergence threshold Maximum number of iterations ;initialization , Lagrange multipliers Punishment factor ;

[0087] 2. Iterative updates:

[0088] No. The next iteration ( ),fixed , ,renew :

[0089] ;

[0090] fixed , ,renew :

[0091] ;

[0092] fixed , ,renew :

[0093] ;

[0094] In the formula, This is the iteration step size; The duration of the training data is in hours;

[0095] 3. Convergence Criterion: Calculate the rate of change of the IMF components between two adjacent iterations:

[0096] ;

[0097] like or If the iteration fails, stop and output the trained VMD model; otherwise, return to step 2 to continue iterating.

[0098] ③ Model application process

[0099] Real-time collected raw load data Input the trained VMD model and decompose it into 5 stationary IMF components. Due to the low calorific value of natural gas This will affect the actual gas consumption of the user's energy-consuming equipment (the higher the calorific value, the lower the gas consumption for the same heat output demand), and the IMF component needs to be corrected. The formula is:

[0100] ;

[0101] In the formula, For the revised first One IMF component (unit: m³ / h); MJ / m³ is the standard lower calorific value of natural gas.

[0102] Calculation example: If m³ / h MJ / m³, then m³ / h.

[0103] (2) Improved Kernel Extreme Learning Machine (IKELM): Load Reference Model Core

[0104] The core function of IKELM is to construct load reference models for different types of users based on the corrected IMF components and gas composition data, and output load reference values. Compared to traditional extreme learning machines, the IKELM scheme in this paper introduces a radial basis kernel function (to improve nonlinear fitting ability) and an L2 regularization term (to avoid overfitting), which are the core guarantees for the accuracy of the reference model.

[0105] ① Model building process

[0106] The formula for IKELM's model output (load reference value) is:

[0107] ;

[0108] In the formula, For the first Reference load values ​​for user class (unit: m³ / h);

[0109] The kernel function mapping matrix is ​​used to map low-dimensional input features to a high-dimensional feature space, thereby linearizing nonlinear relationships. This scheme uses the radial basis function (RBF), and the formula for calculating a single element is as follows:

[0110] ;

[0111] In the formula, Core width (determined through cross-validation); As the input feature matrix, the corrected IMF components, the proportion of gas components, and the corrected lower calorific value are integrated, i.e. ; For the first The input feature vector of each training sample;

[0112] It is an identity matrix (with dimensions consistent with the number of training samples).

[0113] This is the regularization coefficient (initial value, which is adaptively adjusted by the core control layer) used to balance the model's fitting accuracy and generalization ability, and to avoid overfitting.

[0114] The hidden layer output matrix (obtained by kernel function mapping);

[0115] This is a label for actual user energy load (unit: m³ / h), used for error calibration during model training.

[0116] ② Model training process

[0117] The training dataset consists of two parts: one is the IMF components after VMD decomposition and correction. ( (Up to 5), and secondly, synchronous collection. ( Up to 7) and the corresponding actual load label The training sample size consists of 30 consecutive days of user energy consumption data (2880 samples in total, 15 minutes / sample). The training steps are as follows:

[0118] 1. Data partitioning: The dataset is divided into a training set (2016 samples) and a validation set (864 samples) in a 7:3 ratio.

[0119] 2. Kernel function mapping: Inputting the training set into the feature matrix The hidden layer output matrix is ​​mapped using the radial basis kernel function. ;

[0120] 3. Model parameter solution: Solve for the output weight matrix according to the IKELM formula. ;

[0121] 4. Error Calculation: Input the validation set into the feature matrix. Input the model and calculate the prediction error. ( To verify the actual load of the set and calculate the error percentage ;

[0122] 5. Adaptive correction: If If the prediction error exceeds 5%, it indicates that the current VMD decomposition level is [not specified]. Not fully adapted to load fluctuation characteristics, adjustments are needed. (maximum The original load data was re-decomposed using VMD, and the new data was then... Input IKELM and train again until... ;

[0123] 6. Model Saving: Save the trained IKELM model (including the output weight matrix). nuclear width Regularization coefficient Saved for real-time prediction.

[0124] ③ Model application process

[0125] Real-time data collection ( Up to 5) ( Up to 7) Input the trained IKELM model and output the 24-hour load reference values ​​for different user categories. Simultaneously, the correction coefficients are combined with feedback from the core control layer. Output the final load reference value:

[0126] ;

[0127] In the formula, This is the final output load reference value (unit: m³ / h). This is the load reference value correction factor. A positive number indicates that the predicted load is too low and needs to be adjusted upwards, while a negative number indicates that the predicted load is too high and needs to be adjusted downwards.

[0128] Calculation example: If m³ / h (residential users) ,but m³ / h.

[0129] (3) Thermal efficiency reference model

[0130] Based on load reference values ​​and gas characteristics, a thermal efficiency reference model is constructed for different types of users, and thermal efficiency reference values ​​are output. The formula is:

[0131] ;

[0132] In the formula, For the first Thermal efficiency reference values ​​for user classes (unit: %)

[0133] For the first Standard thermal efficiency of energy-consuming equipment for different user categories (90% for residential stoves, 88% for commercial catering stoves, and 92% for industrial kilns).

[0134] For the first The historical average energy load (unit: m³ / h) for this user type is calculated from the data collected by the smart gas meter over the past 30 days, using the following formula:

[0135] ;

[0136] In the formula, The total number of samples over 30 days; For the first The actual load of each sample.

[0137] Calculation example: If (Residential users) MJ / m³ m³ / h m³ / h, then:

[0138] .

[0139] 3. Output and cross-module integration

[0140] The module ultimately outputs two types of data:

[0141] Reference values ​​for final load of different user types in 24 hours ( (to 3)

[0142] Thermal efficiency reference values ​​for different user types ( Up to 3).

[0143] The aforementioned data is synchronously transmitted to the core control layer as the benchmark target value for adaptive control; simultaneously, the module receives feedback from the core control layer. This enables real-time correction of the reference model, forming a sub-closed loop of reference model output - core control layer correction - reference model optimization.

[0144] (III) Core Control Layer: Adaptive Control and Parameter Optimization Module

[0145] refer to Figure 3 This module is the core of the solution, integrating the Gravity Search Algorithm (GSA) and the Fuzzy Adaptive Control Algorithm (FAC) to realize a complete adaptive control logic of global parameter optimization, real-time deviation correction, and parameter self-tuning.

[0146] 1. Input data parsing

[0147] The input data covers the core output of the preceding module:

[0148] Reference model layer output: ( up to 3), ( (to 3)

[0149] Perception layer output: ( Up to 7) ;

[0150] Execution level feedback: Pressure at each level of the pipeline network ( Main trunk line For branch lines, For branch lines, For user access line), pressure after voltage stabilization ( 3) Pressure energy recovery power generation ;

[0151] Real-time deviation data: load deviation thermal efficiency deviation ( (This refers to the average thermal efficiency).

[0152] 2. Core Algorithm Design

[0153] (1) Gravity Search Algorithm (GSA): Global Control Parameter Optimization

[0154] The core function of GSA is to simulate the gravitational interactions between celestial bodies, achieving multi-objective global optimization through the mutual attraction between particles, and determining the initial optimal values ​​of control parameters. This addresses the problem that traditional optimization algorithms are prone to getting trapped in local optima.

[0155] ① Model building process

[0156] Step 1: Define the optimization objective (fitness function). System energy efficiency optimization is a multi-objective problem, requiring the simultaneous achievement of maximum user thermal efficiency, maximum pressure energy recovery power generation, and minimum pipeline transmission and distribution energy consumption. The fitness function formula is:

[0157] ;

[0158] In the formula, The fitness value (a higher value indicates better system energy efficiency);

[0159] , , The weighting coefficients are determined using the analytic hierarchy process (AHP) to ensure a balanced optimization of thermal efficiency, power generation, and energy consumption.

[0160] The average thermal efficiency (in %) for all types of users is given by the formula: ;

[0161] The pressure energy recovery power generation (unit: kWh) is fed back from the execution layer;

[0162] The energy consumption for pipeline transmission and distribution (unit: kWh) is calculated using the following formula:

[0163] ;

[0164] In the formula, Pipeline impedance (unit: MPa·s / m³). The time step is 15 minutes (hours). This is the energy conversion coefficient (calibrated by the pipeline material and flow characteristics).

[0165] Step 2: Define optimization variables. The optimization variables for GSA are the core control parameters affecting system energy efficiency, including: the number of VMD decomposition layers. IKELM regularization coefficient Pipeline impedance Turbine speed That is, optimizing the variable vector .

[0166] Step 3: Construct the core formula system for gravity search to update the particle position (corresponding to the optimization variable):

[0167] Gravity calculation formula:

[0168] ;

[0169] In the formula, For the first The particle pair Gravity of individual particles (characterizing the mutual influence between particles). The gravitational constant (decays with the number of iterations, achieving global search in the early stage and local optimization in the later stage), is given by the formula: ( Let be the initial gravitational constant. The attenuation coefficient is... (Maximum number of iterations); , The first The active mass of the first particle is the same as that of the second particle. The passive mass of each particle (all positively correlated with the particle's fitness value) is given by the formula: ( For the first The fitness value of each particle. , (the maximum / minimum fitness value in the current iteration). For particles and The Euclidean distance; To be the minimum value (to avoid the denominator being zero);

[0170] Acceleration calculation formula:

[0171] ;

[0172] In the formula, For the first The acceleration of each particle; Use random numbers (to increase the randomness of the search); For the first The total mass of the particles ( , (Number of particles);

[0173] Speed-updating:

[0174] ;

[0175] In the formula, For the first The updated velocity of each particle; It is a random number; For the first The current velocity of each particle;

[0176] Position update format:

[0177] ;

[0178] In the formula, For the first The updated position of each particle (i.e., the updated value of the optimization variable).

[0179] ② Model training process

[0180] The training dataset consists of historical data from system operation (including all the input data mentioned above), and the initial range of the optimization variables is: , , MPa·s / m³ r / min. The training steps are as follows:

[0181] 1. Particle initialization: randomly generated There are 10 particles, each corresponding to a set of optimization variables. ( (up to 50)

[0182] 2. Fitness Calculation: Substitute the optimization variables for each particle into the fitness function to calculate... ;

[0183] 3. Iterative optimization:

[0184] No. The next iteration ( ), calculate the gravitational constant Particle mass , ;

[0185] Calculate interparticle gravity acceleration ,speed ,Location ;

[0186] Calculate the fitness value of the updated particles. ;

[0187] 4. Convergence Criterion: If ( (The optimal fitness value for the current iteration) or If the iteration fails, stop; otherwise, return to step 3 and continue iterating.

[0188] 5. Optimal solution output: Outputs the position of the particle with the highest fitness value. That is, the globally optimal control parameters.

[0189] ③ Model application process

[0190] By substituting real-time input data into the trained GSA model, the globally optimal control parameters under the current operating conditions are calculated. This parameter will be used as input to the FAC for real-time deviation correction.

[0191] (2) Fuzzy Adaptive Control Algorithm (FAC): Real-time Deviation Correction and Parameter Self-tuning

[0192] The core function of FAC is based on real-time system deviation (load deviation). thermal efficiency deviation The globally optimal parameters output by GSA are corrected in real time, and adaptive correction coefficients are output. , This achieves collaborative optimization of global optimum and real-time fine-tuning.

[0193] ① Model building process

[0194] Step 1: Define fuzzy input / output variables.

[0195] Input variable (real-time deviation):

[0196] Load deviation The domain of discourse is The corresponding fuzzy linguistic variables are small, medium, and large.

[0197] Thermal efficiency deviation The domain of discourse is The corresponding fuzzy linguistic variables are small, medium, and large.

[0198] Output variables (adaptive correction parameters):

[0199] Load reference value correction factor The domain of discourse is The corresponding fuzzy linguistic variables are negative (large) and zero (large) or positive (large);

[0200] GSA Gravitational Constant Correction Value The domain of discourse is The corresponding fuzzy linguistic variables are negative (large), zero (large), and positive (large).

[0201] Step 2: Define the membership function. All variables use trigonometric membership functions (simple to calculate and highly adaptable), based on load deviation. For example, the membership functions for small, medium, and large are:

[0202] ;

[0203] ;

[0204] ;

[0205] Step 3: Build a fuzzy rule base. Example rule base: 9 rules (covering all input combinations):

[0206] 1. If Small and Small, then , ;

[0207] 2. If Small and In the middle, then , ;

[0208] 3. If Small and Large, then , ;

[0209] 4. If and Small, then , ;

[0210] 5. If and In the middle, then , ;

[0211] 6. If and Large, then , ;

[0212] 7. If large and Small, then , ;

[0213] 8. If large and In the middle, then , ;

[0214] 9. If large and Large, then , .

[0215] Step 4: Defuzzification. The centroid method is used to convert the fuzzy output into a precise value. The formula is:

[0216] ;

[0217] ;

[0218] In the formula, For the first The membership degree of a fuzzy rule (calculated from the membership function of the input variable); , The first The output value corresponding to each rule.

[0219] ② Model training process

[0220] The training dataset is the output of GSA. Deviation from system in real time , The training objective is to minimize the output error after deblurring (the deviation between the correction parameters and the actual optimal correction value). The training steps are as follows:

[0221] 1. Dataset Construction: Collect 1000 sets , With the corresponding optimal correction value ( , );

[0222] 2. Membership function adjustment: Initialize the vertex parameters of the membership function and calculate the fuzzy output for each set of data. , );

[0223] 3. Error Calculation: Calculate the output error. , ;

[0224] 4. Parameter optimization: Gradient descent is used to adjust the vertex parameters of the membership function to minimize the total error. ( (sample size)

[0225] 5. Model saving: Save the trained FAC model (including membership function parameters and fuzzy rule base).

[0226] ③ Model application process

[0227] Real-time data collection , Input the trained FAC model, output and :

[0228] Feedback is sent to the reference model layer to correct the load reference value of IKELM;

[0229] Feedback is sent to the GSA algorithm to adjust the gravitational constant. This will improve GSA's real-time response capabilities.

[0230] At the same time, the output of GSA Transformed into control commands for each module:

[0231] 1. Reference model layer: IKELM regularization coefficients are updated to ;

[0232] 2. Execution layer: Pipeline impedance adjusted to... Turbine speed adjusted to ;

[0233] 3. Perception layer: Sampling frequency (times / minute).

[0234] (3) Stability verification of adaptive control energy-saving system

[0235] To ensure the global asymptotic stability of the adaptive control energy-saving system, a Lyapunov function is designed:

[0236] ;

[0237] In the formula, This is the load error; This is for thermal efficiency error; For parameter estimation error ( For optimal parameters, (For estimating parameters). It is a positive definite matrix.

[0238] right Differentiate:

[0239] ;

[0240] By designing adaptive control laws, ensure ( , This proves that the system is globally asymptotically stable, meeting the core requirements for the stability of control systems.

[0241] 3. Output and cross-module integration

[0242] The module ultimately outputs two types of core content:

[0243] Adaptive control parameters: (Feedback to the reference model layer) (Feedback to GSA) , , , (Transmitted to the execution layer);

[0244] Control commands for each module: Sampling frequency of the sensing layer Reference model layer regularization coefficients Execution layer impedance / speed parameters.

[0245] The above output ensures that the core control layer forms a closed loop with all other modules, realizing a complete adaptive control logic of global optimization, real-time correction, and parameter self-tuning.

[0246] (iv) Execution layer: Transmission and distribution network impedance regulation and pressure energy recovery module (adaptive control actuator)

[0247] refer to Figure 4 This module, as the execution end of the adaptive control energy-saving system, strictly follows the control parameters output by the core control layer, adjusts physical parameters such as pipeline impedance and turbine speed, and transforms the adaptive control commands into the actual pipeline operating state.

[0248] 1. Input data parsing

[0249] The input data consists of control parameters output by the core control layer:

[0250] (Target impedance of pipeline network, unit: MPa·s / m³);

[0251] (Target turbine speed, unit: r / min);

[0252] (Load reference value, unit: m³ / h);

[0253] (Reference value for thermal efficiency, unit: %)

[0254] (User demand pressure, unit: MPa, residential 0.002, commercial 0.003, industrial 0.005).

[0255] 2. Core Processing Procedure

[0256] (1) Adaptive adjustment of pipeline impedance

[0257] Step 1: Calculate the optimal pressure range for the pipeline network This ensures that the pipeline pressure loss is minimized while meeting the user's thermal efficiency requirements.

[0258] Minimum pressure Calculation formula:

[0259] ;

[0260] In the formula, For the first Average thermal efficiency for user class; 0.1 MPa is the minimum safe pressure threshold for the pipeline network;

[0261] Calculation example: If MPa (residential users) ,but MPa;

[0262] Maximum pressure Calculation formula:

[0263] ;

[0264] In the formula, 0.05MPa is the optimal pressure redundancy value (verified by safety and energy consumption balance).

[0265] Calculation example: MPa.

[0266] Step 2: Verify the target impedance of the pipeline network The rationality is demonstrated by the formula:

[0267] ;

[0268] Calculation example: If m³ / h, then MPa·s / m³.

[0269] Step 3: Adjust the pipeline impedance using the variable impedance adjustment unit to... The regulating unit consists of an intelligent throttle valve, an adjustable damping ring, and a flexible inner wall bushing. The regulating formulas for each component are as follows:

[0270] Intelligent throttle valve opening:

[0271] ;

[0272] In the formula, Valve opening degree (0-100%); 0.1 MPa·s / m³ is the maximum allowable impedance of the pipeline network;

[0273] Calculation example: ;

[0274] Adjustable damping ring diameter:

[0275] ;

[0276] In the formula, Ring diameter (unit: mm);

[0277] Calculation example: mm;

[0278] Roughness of flexible inner wall bushing:

[0279] ;

[0280] In the formula, Roughness (unit: μm);

[0281] Calculation example: μm.

[0282] (2) Pressure energy recovery and pressure stabilization adaptive control

[0283] Step 1: Calculate the power generation from pressure energy recovery The formula is:

[0284] ;

[0285] In the formula, The energy recovery coefficient is calculated by multiplying the turbine mechanical efficiency by 90% and the generator electrical efficiency by 95%. The hour is the time step;

[0286] Calculation example: If MPa (main line) MPa m³ / h, then kWh.

[0287] Step 2: Adjust the turbine speed to... The formula is:

[0288] ;

[0289] In the formula, The actual turbine speed (unit: r / min); 0.1 MPa is the reference pressure difference;

[0290] Calculation example: r / min (i.e.) (r / min).

[0291] Step 3: Precise voltage regulation control to ensure output pressure MPa, the formula for the opening degree of the pressure regulating valve is:

[0292] ;

[0293] In the formula, For the pressure regulating valve opening (0-100%);

[0294] Calculation example: (If it exceeds 100%, take 100%).

[0295] 3. Output and cross-module integration

[0296] The module outputs two types of data:

[0297] Operational status data: Pressure at each level of the pipeline network , Recovering electricity Pressure after stabilization ;

[0298] Control execution result: Intelligent throttle valve opening Adjustable damping ring diameter Roughness of flexible inner wall bushing Turbine speed .

[0299] The above data is synchronously fed back to the core control layer to calculate real-time deviations, optimize adaptive control parameters, and form an execution closed loop of core control layer instructions - execution layer actions - core control layer corrections.

[0300] (v) Feedback layer: End-to-end data acquisition and transmission module (key support for adaptive closed loop)

[0301] The core function of this module is to collect operational data from the entire system chain, calculate real-time deviations, and transmit them back to the core control layer to ensure the real-time performance and accuracy of adaptive control.

[0302] 1. Data Collection Standards

[0303] The collected data is as follows:

[0304] Perception layer: , , , , ;

[0305] Reference model layer: , ;

[0306] Execution layer: , , , , , , ;

[0307] Actual operating data: actual user load Actual thermal efficiency .

[0308] The data acquisition frequency is once every 15 minutes (synchronized with the control cycle). Data transmission uses industrial Ethernet (Modbus TCP / IP protocol), with a transmission delay of <100ms. Data verification uses the CRC32 algorithm to ensure that there is no data loss or error.

[0309] 2. Deviation Calculation and Feedback

[0310] Calculate two types of core deviations and feed them back to the core control layer:

[0311] Load deviation ;

[0312] Thermal efficiency deviation .

[0313] Deviation data is transmitted to the core control layer in real time, serving as input to the FAC and driving the real-time correction of adaptive control parameters.

[0314] IV. Specific Implementation Steps (Complete Process of Implementing Adaptive Control Energy-Saving System)

[0315] Step 1: System Deployment and Initialization (Hardware + Algorithm)

[0316] 1. Hardware deployment:

[0317] Sensing layer: Install gas chromatography-mass spectrometry and temperature / humidity / atmospheric pressure sensors at the city gate station;

[0318] Reference model layer / core control layer: Deploy an industrial server (CPU: Intel Xeon Gold 6348, memory: 64GB, hard disk: 1TB SSD), and install Python 3.9 + TensorFlow 2.8 runtime environment;

[0319] Execution layer: Install intelligent throttling valves, adjustable damping rings, and flexible inner wall bushings on the main / branch lines of the pipeline network, and deploy micro turbine generator voltage stabilizers in the regional pressure regulating stations;

[0320] Feedback layer: Pressure sensors (accuracy ±0.0001MPa) are installed at each level of the pipeline network, and smart gas meters are installed at the user end;

[0321] Network deployment: Build an industrial Ethernet network covering all modules with a transmission latency of <100ms.

[0322] 2. Algorithm initialization:

[0323] Initialize VMD ( , ), IKELM , ), GSA ( , ), FAC (Fuzzy Rule Base Initial Parameters);

[0324] Collect 72 hours of basic data to build the initial training dataset.

[0325] Step 2: Algorithm model training (layered training + coupling calibration)

[0326] 1. Reference model layer training:

[0327] Train the VMD model, decompose 72-hour load data, and adjust... To an error ≤ 5%;

[0328] Train the IKELM model by inputting the corrected IMF components and gas composition data, and optimize the regularization coefficient. ;

[0329] 2. Core control layer training:

[0330] Train the GSA model, iterate 100 times, and output the initial value. ;

[0331] Train the FAC model, adjust the membership function parameters, and minimize the output error;

[0332] 3. Coupling calibration: This involves calibrating the FAC output... Feedback is sent to IKELM to correct the load reference value and complete the initial algorithm coupling.

[0333] Step 3: Adaptive closed-loop operation (15 minutes / cycle)

[0334] 1. Data Acquisition: Data acquisition at the perception layer , Feedback layer acquisition , ;

[0335] 2. Reference Model Output: VMD decomposes load data → IKELM output →Combination Output ;

[0336] 3. Adaptive control decision-making: GSA calculation Value and →FAC combination , Output , ;

[0337] 4. Execution layer action: Adjust the pipeline impedance to... Adjust the turbine speed to This enables pressure energy recovery and pressure stabilization.

[0338] 5. Feedback Correction: Feedback layer data collection , , ,calculate , It is then transmitted back to the core control layer.

[0339] Step 4: Algorithm Iteration and Optimization (Regular + Real-time)

[0340] 1. Regular optimization:

[0341] Weekly: Retrain GSA / FAC, update the fuzzy rule base and gravity coefficient decay parameters;

[0342] Monthly: Retrain VMD / IKELM and adjust according to seasonal load changes. , ;

[0343] Quarterly: System stability is checked using Lyapunov functions to ensure the effectiveness of adaptive control.

[0344] 2. Real-time optimization: When or When this happens, an emergency optimization is triggered, the FAC model is retrained, and the corrected parameters are output.

[0345] V. System Innovation and Core Value

[0346] (I) Core Innovation

[0347] 1. Innovative Adaptive Control Framework: For the first time, the GSA+FAC fusion algorithm is applied to the field of gas transmission and distribution, constructing an adaptive control system of global optimization + real-time correction, breaking through the limitations of traditional fixed parameter control;

[0348] 2. Multivariable Cooperative Adaptive: Simultaneously optimizes four types of parameters: load forecasting, pipeline impedance, turbine speed, and sampling frequency. Through FAC, it achieves cooperative correction of multivariable deviations, improving control accuracy to over 98%.

[0349] 3. Innovative Stability Guarantee: The introduction of Lyapunov function design ensures global asymptotic stability of the system, meeting the stringent requirements of industrial adaptive control energy-saving systems.

[0350] (II) Core Values

[0351] 1. Energy efficiency optimization: Pipeline transmission and distribution energy consumption is reduced by 20%-25%, pressure energy recovery power generation is increased by 15%-20%, and user thermal efficiency is stabilized at 93%-96%;

[0352] 2. Enhanced robustness: Adaptable to fluctuations in fuel composition (±10%) and sudden load changes (±30%), requiring no manual intervention;

[0353] 3. Improved economic efficiency: The daily power generation of a single voltage regulating station is ≥200kWh, which can fully cover the daily power consumption of the voltage regulating station and save ≥50,000 yuan in electricity costs per year.

[0354] The embodiments described above are for illustrative purposes only and are not intended to limit the invention. Therefore, any changes in numerical values ​​or substitutions of equivalent elements should still fall within the scope of this invention.

[0355] The above detailed description will enable those skilled in the art to understand that the present invention can indeed achieve the aforementioned objectives and has complied with the provisions of the Patent Law.

[0356] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention. The above descriptions are merely preferred embodiments of the invention and are not intended to limit the invention. It should be noted that any modifications, equivalent substitutions, and improvements made within the spirit and principles of the invention should be included within the scope of protection of the invention.

[0357] It should be noted that the above description of the process is for illustrative purposes only and does not limit the scope of this specification. Those skilled in the art can make various modifications and changes to the process under the guidance of this specification. However, these modifications and changes remain within the scope of this specification.

[0358] The basic concepts have been described above. Obviously, for those skilled in the art who have read this application, the above disclosure is merely illustrative and does not constitute a limitation of this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, and therefore, such modifications, improvements, and corrections still fall within the spirit and scope of the exemplary embodiments of this application.

[0359] Furthermore, this application uses specific terms to describe embodiments of the application. For example, "an embodiment," "an embodiment," and / or "a number of embodiments" refer to a particular feature, structure, or characteristic related to at least one embodiment of the application. Therefore, it should be emphasized and noted that an embodiment, an embodiment, or an alternative embodiment mentioned twice or more in different locations in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of the application can be appropriately combined.

[0360] Furthermore, those skilled in the art will understand that aspects of this application can be described and illustrated through several patentable types or situations, including any new and useful combination of processes, machines, products, or substances, or any new and useful improvements thereof. Therefore, aspects of this application can be implemented entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or a combination of hardware and software. All of the above hardware or software can be referred to as units, modules, or systems. Furthermore, aspects of this application can take the form of a computer program product embodied in one or more computer-readable media, wherein computer-readable program code is contained therein.

[0361] The computer program code required for the operation of each part of this application can be written in any one or more programming languages, including object-oriented programming languages ​​such as Java, Scala, Smalltalk, Eiffel, JADE, Emerald, C++, C#, VB.NET, and Python; general programming languages ​​such as C; Visual Basic, Fortran2103, Perl, COBOL2102, PHP, and ABAP; dynamic programming languages ​​such as Python, Ruby, and Groovy; or other programming languages. This program code can run entirely on the user's computer, or as a standalone software package on the user's computer, or partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer can be connected to the user's computer via any network, such as a local area network (LAN) or wide area network (WAN), or connected to an external computer (e.g., via the Internet), or in a cloud computing environment, or used as a service such as Software as a Service (SaaS).

[0362] Furthermore, unless expressly stated in the claims, the order of processing elements and sequences, the use of numbers and letters, or other names described in this application are not intended to limit the order of the processes and methods of this application. Although some currently considered useful embodiments of the invention have been discussed in the foregoing disclosure by way of various examples, it should be understood that such details are for illustrative purposes only, and the appended claims are not limited to the disclosed embodiments; rather, the claims are intended to cover all modifications and equivalent combinations that conform to the substance and scope of the embodiments of this application. For example, although the implementation of the various components described above can be embodied in a hardware device, it can also be implemented as a purely software solution, such as an installation on an existing server or mobile device.

[0363] Similarly, it should be noted that, in order to simplify the description of the present application and thus aid in the understanding of one or more embodiments of the invention, the foregoing description of the embodiments of the present application sometimes combines multiple features into a single embodiment, drawing, or description thereof. However, this approach of the present application should not be construed as reflecting an intention that the claimed subject matter requires more features than expressly recited in each claim. Rather, the subject of the invention should possess fewer features than in any single embodiment described above.

Claims

1. A municipal gas transmission and distribution integrated energy-saving system, characterized in that, The system comprises a perception layer, a reference model layer, a core control layer, an execution layer, and a feedback layer, which interact sequentially, forming a closed-loop data link between each layer. The perception layer collects gas composition data and environmental parameter data, which are then transmitted to the reference model layer as input data. The reference model layer integrates variational mode decomposition (VMD) and an improved kernel extreme learning machine (KEM) algorithm. Based on the input data, the reference model layer outputs categorized user load reference values ​​and thermal efficiency reference values, which are transmitted to the core control layer as baseline input data. The core control layer integrates a gravity search algorithm and a fuzzy logic algorithm. The adaptive control algorithm involves the core control layer outputting adaptive control parameters based on baseline input data and deviation data transmitted from the feedback layer. These adaptive control parameters are transmitted to the reference model layer, the sensing layer, and the execution layer. The execution layer adjusts the operating state based on the adaptive control parameters and outputs operating state data. This operating state data is transmitted to the feedback layer, which calculates deviation data based on the operating state data and the user's actual energy consumption data. This deviation data is then transmitted to the core control layer. The user's actual energy consumption data originates from the instantaneous gas consumption, cumulative gas consumption, and energy consumption period distribution data collected by smart gas meters deployed at the user's end.

2. The municipal gas transmission and distribution integrated energy-saving system according to claim 1, characterized in that, The sensing layer includes a gas component detection unit and an environmental parameter acquisition unit. The gas component detection unit detects gas components including methane, ethane, propane, n-butane, isobutane, nitrogen, and carbon dioxide. The environmental parameter acquisition unit is a built-in temperature and humidity sensor and an atmospheric pressure sensor. The environmental parameters acquired by the environmental parameter acquisition unit include temperature, humidity, and atmospheric pressure.

3. The municipal gas transmission and distribution integrated energy-saving system according to claim 1, characterized in that, The variational mode decomposition algorithm of the reference model layer decomposes the user's actual energy consumption data into multiple stationary intrinsic mode functions. The improved kernel limit learning machine algorithm uses the intrinsic mode functions, gas composition data, and low calorific value data after environmental parameter correction as input features. The reference model layer is equipped with an error feedback mechanism. When the prediction error of the improved kernel limit learning machine algorithm exceeds a preset threshold, the number of decomposition layers of the variational mode decomposition algorithm is adjusted and the user's actual energy consumption data is decomposed again.

4. The municipal gas transmission and distribution integrated energy-saving system according to claim 1, characterized in that, The gravity search algorithm of the core control layer takes the highest user thermal efficiency, the largest pressure energy recovery power generation, and the lowest pipeline transmission and distribution energy consumption as the multi-objective optimization direction. The optimization variables of the gravity search algorithm include the number of decomposition layers of the variational mode decomposition algorithm, the regularization coefficient of the improved kernel extreme learning machine algorithm, the pipeline impedance value, and the turbine speed value.

5. The municipal gas transmission and distribution integrated energy-saving system according to claim 1, characterized in that, The fuzzy adaptive control algorithm of the core control layer takes load deviation and thermal efficiency deviation as input variables, and load reference value correction coefficient and gravitational constant correction value of gravitational search algorithm as output variables. The fuzzy rule base of the fuzzy adaptive control algorithm contains 9 rules covering all combinations of input variables. The load deviation is the percentage difference between the load reference value and the user's actual energy consumption data, and the thermal efficiency deviation is the percentage difference between the average thermal efficiency and the preset target thermal efficiency.

6. The municipal gas transmission and distribution integrated energy-saving system according to claim 1, characterized in that, The execution layer includes a pipeline impedance adjustment unit and a pressure energy recovery and stabilization unit. The pipeline impedance adjustment unit includes an intelligent throttle valve, an adjustable damping ring, and a flexible inner wall bushing. The pipeline impedance adjustment unit adjusts the opening of the intelligent throttle valve, the diameter of the adjustable damping ring, and the roughness of the flexible inner wall bushing based on the pipeline impedance target value output by the core control layer.

7. The municipal gas transmission and distribution integrated energy-saving system according to claim 6, characterized in that, The pressure energy recovery and stabilization unit adopts a micro turbine generator and voltage stabilization integrated machine. The pressure energy recovery and stabilization unit adjusts the turbine speed based on the turbine speed target value output by the core control layer. The pressure energy recovery and stabilization unit adjusts the opening of the pressure stabilizing valve according to the pipeline pressure data and the user demand pressure data. The user demand pressure data comes from the rated pressure parameters of various types of user energy-consuming equipment.

8. The municipal gas transmission and distribution integrated energy-saving system according to claim 1, characterized in that, The feedback layer includes a full-link data acquisition unit and a deviation calculation unit. The data acquisition unit acquires gas composition data and environmental parameter data from the sensing layer, load reference values ​​and thermal efficiency reference values ​​from the reference model layer, and pipeline pressure data and power generation data from the execution layer. The deviation data calculated by the deviation calculation unit includes load deviation and thermal efficiency deviation.

9. The municipal gas transmission and distribution integrated energy-saving system according to claim 1, characterized in that, The core control layer is equipped with a stability verification unit, which uses a Lyapunov function to verify the global asymptotic stability of the system. The Lyapunov function includes load error variables, thermal efficiency error variables, and parameter estimation error variables.

10. The municipal gas transmission and distribution integrated energy-saving system according to claim 1, characterized in that, The system is equipped with an iterative optimization mechanism, which includes retraining the gravity search algorithm and fuzzy adaptive control algorithm weekly, retraining the variational mode decomposition algorithm and improved kernel extreme learning machine algorithm monthly, and recalibrating the system stability quarterly through a stability verification unit.

Images