Natural gas odorant intelligent control system based on distributed network

Abstract

The invention provides a natural gas odorant intelligent control system based on a distributed network. The natural gas odorant intelligent control system based on the distributed network comprises a sensor monitoring module used for collecting environmental data of natural gas flowing through a pipeline; wherein the environmental data at least comprise pipeline temperature, pipeline pressure, natural gas flow, environmental humidity, gas components in the pipeline and concentration of an odorizor; the intelligent calculation module is used for receiving the environmental data transmitted by the sensor monitoring module through a distributed network, calculating to obtain an odorant injection control instruction based on an odorant adjustment algorithm in combination with the environmental data and historical operation data of the pipeline operation state, and sending the odorant injection control instruction to the odorant control module; the odorizing agent control module is used for adjusting the injection amount of the odorizing agent in the pipeline according to the instruction of the intelligent calculation module so as to adjust the concentration of the odorizing agent in the pipeline; and the remote control module is used for monitoring the operation state of each module so as to enable each module to operate normally.

Description

Technical Field

[0001] The present invention relates to the technical field of control systems, and in particular to a natural gas odorant intelligent control system based on a distributed network. Background Art

[0002] With the widespread use of natural gas, in order to ensure its safe use, odorants need to be added during the production and transportation of natural gas, usually using odorants.

[0003] However, most current natural gas odorization systems rely on manual operation, with outdated monitoring and control methods and a lack of intelligent and automated management. This not only results in waste of odorant but also, in some cases, makes real-time control impossible, leading to natural gas leaks.

[0004] In addition, traditional odorant addition systems have problems such as slow response to environmental factors, complex equipment maintenance, and high operator requirements. Summary of the Invention

[0005] In view of this, an object of an embodiment of the present invention is to provide a natural gas odorant intelligent control system based on a distributed network.

[0006] To achieve the above objectives, the present invention provides a natural gas odorant intelligent control system based on a distributed network, comprising:

[0007] A sensor monitoring module is used to collect environmental data of natural gas flowing through the pipeline; wherein the environmental data at least includes pipeline temperature, pipeline pressure, natural gas flow rate, ambient humidity, gas composition in the pipeline, and odorant concentration;

[0008] an intelligent computing module configured to receive environmental data transmitted by the sensor monitoring module via a distributed network, calculate an odorant injection control instruction based on an odorant adjustment algorithm in combination with the environmental data and historical operating data of the pipeline operation status, and transmit the odorant injection control instruction to the odorant control module; wherein the odorant injection control instruction includes at least an odorant injection amount in the pipeline that needs to be dynamically adjusted;

[0009] an odorant control module, configured to adjust the odorant injection amount in the pipeline according to the odorant injection control instruction, so as to adjust the odorant concentration in the pipeline;

[0010] A remote control module is used to monitor the operating status of the sensor monitoring module, the intelligent computing module and the odorant control module.

[0011] The above technical solution has the following beneficial effects:

[0012] The distributed network-based natural gas odorant intelligent control system includes a sensor monitoring module for collecting environmental data of natural gas flowing through a pipeline; wherein the environmental data includes at least pipeline temperature, pipeline pressure, natural gas flow rate, ambient humidity, gas composition in the pipeline, and odorant concentration; an intelligent computing module for receiving the environmental data transmitted by the sensor monitoring module via the distributed network, and determining the amount of odorant injected into the pipeline based on an odorant adjustment algorithm combined with the environmental data and historical operating data of the pipeline operation status; an odorant control module for adjusting the amount of odorant injected into the pipeline according to instructions from the intelligent computing module to adjust the odorant concentration in the pipeline; and a remote control module for monitoring the operating status of each module to ensure its normal operation. In the present application, the natural gas odorant intelligent control system jointly controls the odorant injection amount through an intelligent computing module and an odorant control module, wherein the intelligent computing module receives the environmental data transmitted by the sensor monitoring module through a distributed network, and determines the odorant injection amount in the pipeline based on the odorant adjustment algorithm combined with the environmental data and the historical operation data of the pipeline operation status. The odorant control module adjusts the odorant injection amount in the pipeline according to the instructions of the intelligent computing module to adjust the odorant concentration in the pipeline. This is conducive to the use of distributed network technology, combined with sensor data acquisition, real-time monitoring, intelligent algorithms and other means to accurately control the odorant injection amount and achieve rapid response to environmental changes. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0014] Figure 1 Schematic diagram of the structure of a natural gas odorant intelligent control system based on a distributed network according to an embodiment of the present invention;

[0015] Figure 2 This is a flow chart of the LSTM model training process according to an embodiment of the present invention;

[0016] Figure 3 is a flow chart of fault diagnosis and automatic adjustment according to an embodiment of the present invention;

[0017] Figure 4 is a flow chart of remote monitoring and management according to an embodiment of the present invention;

[0018] Figure 5 This is a flow chart of a distributed network-based natural gas odorant intelligent control method according to an embodiment of the present invention;

[0019] Figure 6 This is a functional block diagram of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] Example 1

[0022] The present disclosure provides a structural schematic diagram of a natural gas odorant intelligent control system based on a distributed network. Figure 1 FIG is a schematic diagram of the structure of a natural gas odorant intelligent control system based on a distributed network according to an embodiment of the present invention. Figure 1 As shown, the distributed network-based natural gas odorant intelligent control system includes:

[0023] A sensor monitoring module is used to collect environmental data of natural gas flowing through the pipeline; wherein the environmental data at least includes pipeline temperature, pipeline pressure, natural gas flow rate, ambient humidity, gas composition in the pipeline, and odorant concentration;

[0024] an intelligent computing module configured to receive environmental data transmitted by the sensor monitoring module via a distributed network, calculate an odorant injection control instruction based on an odorant adjustment algorithm in combination with the environmental data and historical operating data of the pipeline operation status, and transmit the odorant injection control instruction to the odorant control module; wherein the odorant injection control instruction includes at least an odorant injection amount in the pipeline that requires dynamic adjustment; wherein the historical operating data includes sampled and recorded data of parameters such as natural gas flow, pipeline pressure, ambient temperature and humidity, gas composition, and odorant concentration under different operating conditions, and is used to support the prediction calculation of the odorant injection amount and the generation of an intelligent adjustment strategy;

[0025] an odorant control module, configured to adjust the odorant injection amount in the pipeline according to the odorant injection control instruction, so as to adjust the odorant concentration in the pipeline;

[0026] A remote control module is used to monitor the operating status of the sensor monitoring module, the intelligent computing module and the odorant control module.

[0027] In this exemplary embodiment, the sensor monitoring module may include temperature sensors, pressure sensors, flow sensors, and odorant concentration sensors. These sensors continuously collect environmental data about the natural gas flowing through the pipeline and transmit the real-time data to the intelligent computing module via the data communication module. For example, the temperature sensor collects pipeline temperature, the pressure sensor collects pipeline pressure, the flow sensor collects natural gas flow, and the humidity sensor collects ambient humidity. In this embodiment, gas composition and odorant concentration within the pipeline are acquired by providing a gas composition analysis module, which can be integrated into the sensor monitoring module and deployed at target measurement points within the natural gas pipeline. Specifically, pipeline gas composition can be acquired using equipment such as a micro-gas chromatograph (Micro-GC) or a non-dispersive infrared spectrometer (NDIR). These sensors can perform real-time identification and quantitative analysis of the main components of natural gas (e.g., methane (CH4), ethane (C2H2), carbon dioxide (CO2), nitrogen (N2), etc.), outputting the volume fraction or concentration value of each gas component, thereby obtaining accurate pipeline gas composition data. The detection of odorant concentration can be achieved through a variety of technical means. For example, electrochemical gas sensors or photoionization detectors (PIDs) can be used to specifically detect typical compounds in odorants (such as tetrahydrothiophene (THT) and methyl mercaptan). Pipeline gas sampling and analysis can also be combined with a mass spectrometer (MS) or a Fourier transform infrared spectrometer (FTIR) to accurately quantify the type and concentration of odorants. In scenarios where high response speed is required, laser gas analysis technologies such as tunable diode laser absorption spectroscopy (TDLAS) can also be used to achieve rapid and selective detection of specific odorant molecules. The gas composition analysis module can be set as an online sampling and analysis device with functions such as timed automatic sampling, intelligent calibration, and remote data upload to ensure the accuracy and timeliness of the collected data. The system uploads the obtained gas composition data and odorant concentration data to the intelligent computing module via the data communication module for subsequent dynamic adjustment and optimization control of the odorant injection amount.

[0028] The intelligent computing module can use advanced intelligent algorithms to combine the environmental data transmitted by sensors with the historical operating data of the pipeline operation status. The intelligent computing module can adjust the injection amount of the odorant in real time to ensure that the odorant concentration in each section of the natural gas pipeline meets safety standards.

[0029] In this exemplary embodiment, the distributed network-based natural gas odorant intelligent control system also includes a data communication module. This module utilizes distributed network technology to interconnect various modules, supporting data transmission and remote control, and enabling real-time upload and intelligent feedback of multi-point data.

[0030] In this embodiment, the distributed network refers to a network structure based on the collaborative work of multiple nodes, wherein each sensor node, odorant control unit, gas composition analysis module, data communication module and intelligent computing module, etc., all serve as independent data interaction terminals, and are connected to the same network through wired or wireless communication to achieve information sharing and interaction. The specific implementation can be based on industrial Ethernet (such as Modbus TCP / IP, Profinet), wireless sensor network (WSN), LoRa, NB-IoT or 5G and other communication protocols, and support functions such as multi-point synchronous data acquisition, remote control command issuance and feedback data transmission. The distributed network structure has good scalability and robustness, supports the simultaneous online operation of multiple measurement points and control points, and can realize remote real-time monitoring and scheduling optimization of the odorization status of different pipeline sections in the natural gas transmission and distribution system, thereby improving the intelligence level and responsiveness of the system.

[0031] In this exemplary embodiment, during data collection and processing, all sensor data (e.g., natural gas flow, pressure, temperature, gas composition, etc.) is transmitted to the control center via a distributed network. After receiving the data, the intelligent computing module performs preliminary preprocessing. This stage includes data cleaning, missing value filling, and outlier detection. To ensure data accuracy and reliability, any errors or abnormal data that occur during transmission are flagged and corrected. This preprocessing step ensures the high quality and integrity of the raw data used in the system's subsequent calculations.

[0032] After data preprocessing is completed, the intelligent computing module can input the real-time collected data into the adaptive machine learning algorithm through a dynamic data feedback mechanism. At this stage, the system first integrates the sensor data with the historical operating data of the pipeline operation status. By comparing the data stream and historical records, it identifies the key changing factors of the current environment (such as fluctuations in flow, pressure, temperature and gas composition). Through the feedback mechanism, the system can evaluate and update the adaptive machine learning model in real time, so that it can gradually adapt to different environmental conditions (such as pressure fluctuations in the pipeline or changes in gas composition) and automatically adjust the odorant injection strategy.

[0033] As data continues to flow in, the intelligent computing module activates adaptive machine learning algorithms, such as reinforcement learning or deep reinforcement learning. These algorithms dynamically optimize the odorant injection control strategy based on real-time data and system feedback. The system automatically adjusts to environmental changes, such as pressure fluctuations or changes in gas composition. Rather than relying on fixed, pre-set algorithms, the system continuously learns and optimizes based on new data, ensuring that the system can flexibly adjust injection rates according to real-time needs.

[0034] In this exemplary embodiment, the remote control module can allow operators to remotely access the system through a cloud platform or a dedicated terminal through Internet technology to perform fault diagnosis, system adjustment and maintenance.

[0035] The distributed network-based natural gas odorant intelligent control system in this application combines distributed networks, intelligent computing and real-time monitoring technologies to efficiently and accurately adjust the usage of natural gas odorant to adapt to different environmental changes and user needs, thereby significantly improving the safety and economy of natural gas.

[0036] In some embodiments, the odorant adjustment algorithm comprises:

[0037] Determining a basic odorant injection amount based on the natural gas flow rate, the target odorant concentration, the actual odorant concentration, and the system's automatic increase proportional coefficient;

[0038] determining a pressure correction factor based on the pipeline pressure and the standard pressure;

[0039] determining a pressure-corrected injection amount based on the pressure correction factor and the basic injection amount of the odorant;

[0040] Determining a temperature correction factor based on the pipeline temperature and a standard temperature;

[0041] determining a temperature-corrected injection amount based on the temperature correction factor and the pressure-corrected injection amount;

[0042] determining a composition correction factor based on a methane concentration in the gas composition in the pipeline and a baseline methane concentration;

[0043] determining a component-corrected injection amount based on the component correction factor and the temperature-corrected injection amount;

[0044] determining a humidity correction factor based on the ambient humidity, saturated humidity, and reference humidity;

[0045] The amount of odorant injected into the pipeline is determined based on the humidity correction factor and the component-corrected injection amount; wherein the standard pressure, the standard temperature, the baseline methane concentration, the saturated humidity, and the reference humidity are all historical operating data. The historical operating data in this application includes but is not limited to the above data.

[0046] In this exemplary embodiment, upon system startup, hardware checks and system configuration are first performed. The odorant control module initializes each sensor (temperature sensor, pressure sensor, flow sensor, odorant concentration sensor, etc.) and connects it to the intelligent computing module. Each sensor transmits its initial state data via the data communication module to ensure the system operates within accurate parameters. This allows the system to dynamically adjust its operating state based on real-time collected environmental data, adapting to varying operating environments.

[0047] In this application, in addition to temperature sensors, pressure sensors, flow sensors, and odorant concentration sensors, the following new sensors can also be introduced to improve the intelligence and adaptability of the system. First, gas composition sensors (such as methane sensors) can be used to monitor the concentrations of other gas components in natural gas, such as methane, ethane, etc., to more accurately evaluate the effect of the odorant and optimize the injection amount. Secondly, the environmental humidity sensor can be used to detect changes in humidity in the pipeline or environment. Humidity has an important impact on gas flow and odor diffusion. The addition of this sensor can help the system adjust the odorant injection amount more accurately. In addition, the vibration sensor can be used to monitor the vibration of the odorant injection device and the pipeline system, and promptly detect whether the equipment is operating normally or abnormal, to prevent equipment failure from causing improper odorant injection. Finally, the introduction of an intelligent sensor network, through the joint work of multiple distributed sensors, can achieve more comprehensive and detailed data collection and environmental monitoring, provide more comprehensive support for the intelligent computing module, and ensure the efficiency and accuracy of the odorant injection process. The introduction of these new sensors can effectively improve the intelligence level of the system, enhance its adaptability and fault warning capabilities, and thus facilitate the real-time calculation of the odorant injection amount in the pipeline through the above-mentioned odorant adjustment algorithm.

[0048] In this exemplary embodiment, sensors perform preliminary data preprocessing and filtering through built-in intelligent data processing units, transmitting only high-quality, meaningful raw data. This reduces data redundancy and improves network efficiency. All sensors are connected via a low-power wide-area network, ensuring low latency and efficient transmission even over long distances and with limited bandwidth.

[0049] To ensure real-time and accurate system performance, the system can adopt a distributed network architecture based on edge computing. In this architecture, sensors and data acquisition devices not only send data to a central control system but also perform rapid local edge computing for data pre-analysis and error detection. This approach reduces the burden of data transmission, improves the system's response to emergencies, and promptly identifies potential faults or abnormal data.

[0050] Furthermore, the system utilizes dynamic network optimization technology during data transmission. This technology intelligently adjusts the data transmission path and frequency based on the network status and real-time data load, ensuring that data from each sensor reaches the central control system in real time while maintaining network stability and reliability. This dynamic network optimization technology enables the system to flexibly adapt to complex network environments, ensuring the timely and accurate upload of monitoring data. By integrating gas composition sensors, humidity sensors, and vibration sensors, the system comprehensively monitors the multi-dimensional data of natural gas flowing through the pipeline, improving the system's monitoring accuracy.

[0051] In this exemplary embodiment, edge computing capabilities can be added to the sensor end to pre-process and analyze data, reduce network bandwidth pressure, and improve real-time response speed. Sensors are connected through a low-power network to ensure efficient connection of large-scale distributed sensor systems and reduce energy consumption. Data transmission paths are intelligently adjusted according to network conditions to improve the reliability and flexibility of data transmission. The distributed network-based natural gas odorant intelligent control system of this application not only achieves efficient and accurate real-time data collection and transmission, but also ensures data reliability and response speed in complex and changing network environments, thereby improving the performance and sustainability of the entire natural gas odorant intelligent control system.

[0052] In this exemplary embodiment, data transmitted to the control center is processed by an intelligent computing module. This module analyzes historical operational data based on a preset algorithm and pipeline status, combined with real-time environmental data (e.g., natural gas flow, pressure, and temperature), for a comprehensive assessment. Using this data, the intelligent computing module predicts the optimal odorant injection rate and calculates the amount of odorant to be added to the current pipeline segment. By dynamically analyzing the collected multidimensional data using intelligent algorithms (e.g., machine learning or deep learning algorithms) to calculate the optimal odorant injection rate in real time, this adaptive adjustment capability enhances the system's intelligence and operational efficiency.

[0053] In some embodiments, determining the basic odorant injection amount based on the natural gas flow rate, the target odorant concentration, the actual odorant concentration, and the system automatic increase proportional coefficient includes:

[0054] Among them, f base is the basic injection amount of the odorant; C target is the target odorant concentration, C actual is the actual odorant concentration, Q is the natural gas flow rate, and k1 is the system automatic increase proportional coefficient. The system automatic increase proportional coefficient refers to the odorant intelligent control system detecting the actual odorant concentration C actual Significantly lower than the target odorant concentration C target In order to ensure that the injection concentration catches up with the target value in time, the adjustment proportional factor k1 in the basic injection volume calculation process is automatically amplified dynamically to enhance the sensitivity and response rate of the injection reaction. Among them, the system refers to the natural gas odorant intelligent control system based on a distributed network in this application. The system continuously monitors parameters such as natural gas flow, current injection concentration and concentration deviation through an integrated computing module and sensor network, and calculates the proportional coefficient in real time. The value of this coefficient is dynamically generated by an adjustment function based on concentration deviation, a fuzzy control rule or an adaptive algorithm based on historical operating data of the pipeline operation status. Its main purpose is to quickly increase the odorant injection volume when the gas flow rate changes suddenly, the external environment is disturbed or the concentration error is large, so as to ensure that the final odorant concentration always meets national or industry safety standards.

[0055] In this exemplary embodiment, the amount of odorant injected is positively correlated with the natural gas flow rate. The greater the flow rate, the faster the natural gas velocity in the pipeline, and more odorant needs to be injected to ensure uniform mixing in a short time.

[0056] In terms of dynamic adjustment, the basic odorant injection amount is determined based on the natural gas flow rate, target odorant concentration, actual odorant concentration and the system automatic increase ratio coefficient, including:

[0057] When the flow rate Q exceeds the threshold Q threshold (For example, a sudden increase of more than 20%), the system automatically increases the proportional coefficient k1 to quickly compensate for the demand; if the flow fluctuation standard deviation σ(Q) exceeds the safe range, the moving average and other smoothing algorithms are enabled to adjust the instantaneous injection volume to avoid excessive oscillations.

[0058] In some embodiments, determining the pressure correction factor based on the pipeline pressure and the standard pressure includes:

[0059] Where β(P) is the pressure correction factor, P is the pipeline pressure, and P0 is the standard pressure;

[0060] The step of determining the pressure-corrected injection amount based on the pressure correction factor and the basic injection amount of the odorant includes:

[0061] f pressure =f base β(P); where fpressure is the pressure-corrected injection volume, f base This is the base injection amount of the odorant.

[0062] In this exemplary embodiment, pressure changes will affect the gas volume, thereby changing the odorant concentration per unit volume. When the pressure increases, the injection volume needs to be reduced, and when the pressure decreases, the injection volume needs to be increased. Introducing a pressure correction factor (Based on the standard pressure P_0), the corrected injection volume f pressure =f base β(P).

[0063] Dynamic adjustment includes: when the pressure P>1.2P0, nonlinear corrections such as exponential decay are enabled to prevent excessive injection; when P<0.8P0, additional compensation term Δf=k is added p (P0-P), ensure the coverage of odorant in low-density gas.

[0064] In some embodiments, determining the temperature correction factor based on the pipeline temperature and the standard temperature includes:

[0065] Where γ(T) is the temperature correction factor, T0 is the standard temperature, and T is the pipe temperature;

[0066] The determining of the temperature correction injection amount based on the temperature correction factor and the pressure correction injection amount includes:

[0067] f temp =f pressure γ(T); where f temp is the temperature correction injection amount, f pressure Correct the injection volume for pressure.

[0068] In this exemplary embodiment, temperature and odorant adjustment: temperature increase will cause gas expansion, reduce the concentration of odorant per unit volume, and need to increase the injection amount, and vice versa. Based on the ideal gas law, the temperature correction factor (T0 is the standard temperature, such as 20℃), the corrected injection volume f temp =f pressure γ(T). During dynamic adjustment, when the temperature T>50°C or T<-10°C, an empirical correction coefficient table for quadratic terms fitted with historical operating data is used to compensate for nonlinear effects. At the same time, LSTM is used to predict the temperature trend for the next two hours and adjust the injection amount in advance to avoid lag.

[0069] In some embodiments, determining the composition correction factor based on the methane concentration in the gas component in the pipeline and the baseline methane concentration includes:

[0070] Among them, δ(CCH4 ) is the composition correction factor, C ref is the baseline methane concentration, C CH4 is the methane concentration in the gas components in the pipeline, α is the proportional adjustment coefficient, which reflects the sensitivity of the change in methane concentration to the amount of odorant injected, that is, the sensitivity coefficient of the degree of correction of odorant injection when the component concentration deviates from the reference value;

[0071] The determining of the component correction injection amount based on the component correction factor and the temperature correction injection amount includes:

[0072] f gas =f temp ·δ(C CH4 ); where f gas is the component correction injection amount, f temp Correct the injection volume for temperature.

[0073] In this exemplary embodiment, the methane concentration affects the odorant adsorption efficiency. High-purity methane requires more odorant. Impurity gases (such as CO2 and N2) dilute the odorant and require dynamic compensation. (C ref is the baseline methane concentration), the final injection volume f gas =f temp ·δ(C CH4 ). Dynamic adjustment includes: if the CO2 concentration is detected to be more than 5%, additional compensation Δf CO2 =k CO2 ·C CO2 Q; Use reinforcement learning to train a multi-gas coupling effect model and optimize the injection strategy under complex composition. C CH4 is the actual methane concentration in natural gas, C ref is the baseline methane concentration used to calculate the composition correction factor. CO2 It is the proportional coefficient of carbon dioxide concentration compensation (used to calculate the additional amount of odorant compensation required due to changes in carbon dioxide concentration). CO2 It is the actual concentration of carbon dioxide in natural gas (used to characterize the carbon dioxide content in the gas composition to trigger corresponding odorant compensation adjustment).

[0074] In some embodiments, determining the humidity correction factor based on the ambient humidity, saturated humidity, and reference humidity includes:

[0075] Where η(H) is the humidity correction factor, H sat is the saturated humidity, H is the ambient humidity, H ref is the reference humidity, k H is an adjustable proportional coefficient;

[0076] The step of determining the injection amount of the odorant in the pipeline based on the humidity correction factor and the component correction injection amount includes:

[0077] f humidity =f gas ·η(H); where f humidity is the amount of odorant injected into the pipeline, f gas Correct injection volume for the component.

[0078] In this exemplary embodiment, high humidity will cause odorant molecules to be adsorbed on the pipe wall or in the liquid water film, reducing the effective concentration in the gas phase and requiring an increase in the injection volume. Humidity correction factor (H sat is the saturated humidity, H ref is the reference humidity, for example 50%), the corrected injection volume f humidity =f gas ·η(H). During dynamic adjustment, when the ambient humidity H is greater than 90% and the temperature drops sharply, the anti-condensation mode is triggered, an additional 10% to 15% injection volume is added, and pipeline heating is started. H It is an adjustable proportional coefficient used to quantify the influence of humidity change on the injection amount of odorant. Its physical meaning is: when the actual ambient humidity H deviates from the reference humidity H ref When k H Adjust the magnitude of the correction factor to compensate for the reduction in effective gas phase concentration caused by adsorption of odorant molecules by pipe walls or liquid water films in high humidity environments. H The value of needs to be determined through experiments or simulations based on factors such as pipeline material, gas flow rate, and humidity historical operation data, and is a positive number (the higher the humidity, the greater the compensation amount). gas It is a component correction value that comprehensively considers the impact of natural gas components (such as methane, carbon dioxide, etc.) on the odorant injection amount (used to adjust the odorant injection baseline amount according to the actual gas composition).

[0079] In this exemplary embodiment, the vibration intensity V reflects the operating status of the equipment. Abnormal vibration leads to injection deviation, which requires dynamic correction or maintenance. The adjustment strategy includes: if the vibration acceleration V is greater than the acceleration threshold V threshold , automatically reducing the injection rate and starting redundant pumps; performing spectrum analysis on vibration signals through Fast Fourier Transform (FFT), identifying characteristic frequencies such as bearing wear, predicting equipment life and adjusting maintenance cycles to avoid concentration runaway caused by sudden failures.

[0080] In this embodiment, an intelligent computing module combines advanced adaptive machine learning algorithms (such as reinforcement learning and deep reinforcement learning) with real-time environmental data (such as natural gas flow, pressure, temperature, and gas composition) for comprehensive evaluation. By incorporating a dynamic data feedback mechanism, the intelligent computing module can adjust its model and calculation strategy based on real-time data, enabling online learning. The system adjusts the odorant injection rate in real time based on varying environmental conditions (such as pipeline pressure fluctuations and changes in gas composition), ensuring that the system can adaptively adjust the odorant injection strategy.

[0081] In some embodiments, the odorant control module is specifically configured to:

[0082] Determine the objective function of odorant injection accuracy, resource consumption, equipment load, equipment maintenance cost, environmental friendliness, response speed optimization, and system stability;

[0083] Constructing a comprehensive objective function based on the odorant injection accuracy objective function, resource consumption objective function, equipment load objective function, equipment maintenance cost objective function, environmental friendliness objective function, response speed optimization objective function, and system stability objective function;

[0084] Based on minimizing the comprehensive objective function as the optimization goal, the injection pattern of the odorant in the pipeline is used as the function optimization variable to determine the optimized injection pattern of the odorant;

[0085] Based on the optimized odorant injection pattern, the odorant injection amount in the pipeline is adjusted to adjust the odorant concentration in the pipeline.

[0086] In this exemplary embodiment, in order to improve the accuracy of the odorant injection amount, the intelligent computing module also introduces a multi-objective optimization algorithm, which comprehensively considers the accuracy of the odorant injection amount, resource consumption and equipment load, and performs balanced optimization. Through predictive modeling and error correction mechanisms, the system can predict future environmental changes based on historical operating data of the pipeline operation status and make adjustments in advance, thereby reducing the system's response delay and avoiding odorant concentration fluctuations caused by environmental changes in advance. The optimized odorant injection pattern in this application is determined based on minimizing the comprehensive objective function. For example, it is determined based on a comprehensive consideration of odorant injection accuracy, resource consumption, equipment load, equipment maintenance cost, environmental friendliness, response speed optimization and system stability.

[0087] In some embodiments, the odorant injection accuracy objective function f1(x) is:

[0088] Among them, f actual is the actual injection amount of odorant, f targetTarget injection amount of odorant;

[0089] The resource consumption objective function f2(x) is:

[0090] f2(x)=C energy (x)+C odorant (x); where C odorant (x) is the odorant consumption cost, C energy (x) is the energy consumption cost;

[0091] The equipment load objective function f3(x) is:

[0092] f3(x)=t(x) / L rated ; Among them, t(x) is the equipment load and equipment operation time, L rated is the rated load of the equipment;

[0093] The equipment maintenance cost objective function f4(x) is:

[0094] f4(x)=C maintenance (t(x), I(x)); where t(x) is the equipment load and equipment operation time, and I(x) is the equipment operation intensity coefficient; C maintenance (t(x), I(x)) is a function of running time and running intensity;

[0095] Among them, C maintenance (t(x), I(x)) = k1·t(x) + k2·I(x); k1 and k2 are coefficients related to maintenance costs;

[0096] The environmental friendliness objective function f5(x) is:

[0097] f5(x) = 1 / (V(x)·P(x)); where V(x) is the volatility coefficient of the odorant and P(x) is the amount of harmful byproducts produced;

[0098] The response speed optimization objective function f6(x) is:

[0099] f6(x)=T(x); where T(x) is the time required for the system to complete the adjustment and reach a stable state after receiving the adjustment instruction;

[0100] The system stability objective function f7(x) is:

[0101] f7(x)=1 / σ(x); σ(x) is the standard deviation of the fluctuation of the system odorant injection flow rate;

[0102] The comprehensive objective function F(x) is constructed based on the odorant injection accuracy objective function, resource consumption objective function, equipment load objective function, equipment maintenance cost objective function, environmental friendliness objective function, response speed optimization objective function and system stability objective function, including:

[0103] F(x)=w1·(1-f1(x))+w2·f2(x)+w3·f3(x)+w4·f4(x)+w5·(1-f5(x))+w6·f6(x)+w7·(1-f7(x)); among them, w1, w2, w3, w4, w5, w6 and w7 are the weights corresponding to each objective function; w1+w2+w3+w4+w5+w6+w7=1.

[0104] In this exemplary embodiment, it is desired that the odorant injection accuracy objective function f1(x) is as close to 1 as possible, which can be expressed as To express it, x is the set of decision variables that affect the injection amount, such as equipment parameters, control parameters, etc.

[0105] The resource consumption objective function f2(x) is expected to be as small as possible, and f2(x) is expressed as the energy consumption cost C energy (x) and odorant consumption cost C odorant (x) and the sum of them. Among them, the energy consumption cost C energy (x) is related to the equipment operating power and operating time; the odorant consumption cost C odorant (x) is related to the amount of odorant used and the unit price.

[0106] The equipment load objective function f3(x) is expected to be as small as possible, assuming that the equipment load is related to the equipment operating time t(x) and the equipment rated load L rated Related, can be expressed as: f3(x)=t(x) / L rated .

[0107] The equipment maintenance cost objective function f4(x) is expected to be as small as possible. The equipment maintenance cost is related to factors such as equipment operating time and operating intensity. Assuming that the equipment maintenance cost is related to the equipment operating time t(x) and the equipment operating intensity coefficient I(x) (for example, the degree to which parameters such as pressure and temperature during equipment operation deviate from the normal range), it can be expressed as:

[0108] f4(x)=C maintenance (t(x),I(x)); where C maintenance (t(x), I(x)) is a function of running time and running intensity, for example: C maintenance (t(x),I(x))=k1·t(x)+k2·I(x).

[0109] It is expected that the environmental friendliness objective function f5(x) is as large as possible. It is assumed that environmental friendliness is mainly related to the volatility of the odorant and the amount of harmful by-products produced. Harmful by-products refer to non-target substances produced by physical or chemical reactions during the injection and diffusion of the odorant, which may cause harm to the environment or human health. This system obtains its production amount P(x) through theoretical modeling (establishing a mathematical expression for by-product generation based on reaction mechanism, temperature, pressure and other parameters), historical operation data fitting (using existing experimental or operation data to fit the empirical relationship between by-product output and odorization conditions) or sensor detection (using gas analysis sensors to collect by-product concentrations in real time and calculate the production amount), etc., as the input variable in the environmental friendliness objective function, which is used to comprehensively evaluate the impact of the odorization scheme on the environment. Assuming the volatility coefficient of the odorant is V(x) (the stronger the volatility, the larger the coefficient), and the amount of harmful by-products produced is P(x), the environmental friendliness objective function can be expressed as:

[0110] f5(x)=1 / (V(x)·P(x)); or it can be defined by a more complex environmental impact assessment model.

[0111] The desired response speed optimization objective function f6(x) is as small as possible. Response speed can be measured by the time T(x) it takes for the system to complete the adjustment and reach a stable state after receiving the adjustment instruction, i.e., f6(x) = T(x).

[0112] It is expected that the system stability objective function f7(x) will be as large as possible. System stability can be measured by the degree of fluctuation during system operation. Assuming that the standard deviation of fluctuation of a key operating parameter of the system (e.g., odorant injection flow rate) is σ(x), the system stability objective function can be expressed as: f7(x) = 1 / σ(x);

[0113] The multi-objective optimization problem requires combining these objective functions and using the weighted summation method to construct a comprehensive objective function F(x):

[0114] F(x)=w1·(1-f1(x))+w2·f2(x)+w3·f3(x)+w4·f4(x)+w5·(1-f5(x))+w6·f6(x)+w7·(1-f7(x));

[0115] Among them, w1, w2, w3, w4, w5, w6 and w7 are the weights corresponding to each objective function; w1+w2+w3+w4+w5+w6+w7=1, and w1, w2, w3, w4, w5, w6 and w7 are all greater than or equal to 0.

[0116] These weight coefficients reflect the relative importance of different objectives and are adjusted according to actual conditions (such as safety standards, cost constraints, environmental protection requirements, etc.). The appropriate x is obtained through optimization to minimize F(x), that is, min F(x).

[0117] At the same time, in practical applications, some constraints also need to be considered, such as safety standard restrictions (such as the lower limit of odorant concentration) and the range of equipment operating parameters. Assume that there are safety standard constraints g1(x) ≥ 0 (for example, the odorant concentration must meet the minimum safety concentration requirement, and the olfactory recognition concentration of the odorant throughout the gas transmission process must not be lower than the set threshold), and equipment operating parameter constraints g2(x) ≤ 0 (for example, the equipment operating power cannot exceed the rated power, the injection frequency cannot exceed the safety frequency limit, and the operating temperature, current, pressure, etc. cannot exceed the rated upper limit).

[0118] Specifically, for the safety standard constraint function (safety lower limit of odorant concentration), let the actual odorant concentration in the pipeline be C odor (X), the minimum safe concentration is C min , then g1(x)=C odor (X)-C min , g1(x)≥0. This inequality ensures that the odorant injection pattern must ensure that its concentration reaches the minimum safety concentration requirement.

[0119] Specifically, for the device operating parameter constraint function (such as power upper limit), the actual operating power of the device is P(X), and the maximum allowed power is P max , then g2(x)=P(X)-P max , g2(x)≤0, which ensures that the equipment will not be overloaded.

[0120] In this application, the multi-objective optimization framework indirectly controls the injection volume by comprehensively weighing objectives such as odorant injection accuracy, resource consumption, and equipment load. For example, the injection accuracy objective function requires the actual injection volume to be as close to the theoretical value as possible, which drives the system to dynamically adjust the valve opening according to real-time data of flow and pressure; while the resource consumption objective limits the excessive use of energy and odorants, forcing the system to give priority to a smooth control strategy (such as reducing the pump frequency rather than frequent starting and stopping) when the flow fluctuates. These objectives act synergistically on the decision variable x (such as control parameters, equipment operation mode) through weight coefficients, ultimately affecting the calculation logic of the injection volume.

[0121] Although the objective function doesn't directly include parameters like flow rate and pressure, these parameters indirectly influence the optimization results through decision variables. For example, the operating time t(x) in the equipment load objective is positively correlated with flow rate. This means that at high flow rates, the pump must run continuously, increasing the load. The system lowers the flow threshold to extend equipment life, thereby reducing injection frequency. The volatility coefficient V(x) in the environmental friendliness objective is affected by temperature. At high temperatures, the system needs to reduce the use of highly volatile odorants and dynamically adjust the temperature correction factor to balance concentration and environmental requirements.

[0122] Each objective function is dynamically coordinated through weighting coefficients. For example, when the safety weight (w1) is high, the system tolerates higher energy consumption to ensure concentration compliance, while when the cost weight (w2) is high, slight concentration deviations are tolerated to reduce resource consumption. When equipment maintenance goals conflict with response speed, high-load equipment triggers reduced speed operation. The system adjusts valve openings to compensate for response delays, achieving a balance between maintenance cost and real-time performance. This dynamic trade-off enables the system to maintain global optimization under complex operating conditions.

[0123] In some embodiments, the odorant control module is specifically configured to:

[0124] Training a Long Short-Term Memory (LSTM) network original model based on historical operation data of the pipeline operation status to obtain an LSTM prediction model;

[0125] Predicting future pipeline natural gas flow and pipeline pressure trends based on the LSTM prediction model;

[0126] Based on the optimized odorant injection pattern and in conjunction with future natural gas flow and pipeline pressure trends, the odorant injection rate is adjusted to adjust the odorant concentration within the pipeline. Historical operational data, including sampled and recorded data on parameters such as natural gas flow, pipeline pressure, ambient temperature and humidity, gas composition, and odorant concentration under different operating conditions, is used to support the prediction and calculation of odorant injection rates and the generation of intelligent adjustment strategies.

[0127] In this exemplary embodiment, based on time-series data analysis, the system can also use algorithms such as long-short-term memory (LSTM) networks to predict flow and pressure trends in different pipeline sections over a period of time. This allows for more accurate pre-setting of odorant injection levels and proactive adjustments to address future changes. This approach not only enables the system to respond to current environmental changes in real time but also predicts and adapts to future fluctuations. By introducing reinforcement learning and deep reinforcement learning, the system is able to learn and optimize odorant injection strategies in real time, rather than relying solely on static, pre-set algorithms. Reinforcement learning enables the system to dynamically optimize based on real-time data in the face of a constantly changing environment.

[0128] The intelligent computing module continuously adjusts its algorithm model through real-time data collection and performs online learning. This adaptive learning process not only enhances the system's flexibility but also enables it to continuously optimize its performance over time.

[0129] By combining multiple objectives, such as odorant injection accuracy, resource consumption, and equipment load, multi-dimensional optimization can improve the overall efficiency of the system and reduce excessive consumption.

[0130] Using long short-term memory networks (LSTM) to process and predict time series data can accurately predict environmental changes in the future, make odorant injection adjustments in advance, and reduce the impact of environmental fluctuations.

[0131] By predicting future environmental changes, the system can proactively correct potential injection errors, reducing response delays and improving stability. This not only makes the system more intelligent and adaptive, but also allows it to maintain efficient and precise operation in changing environments, enhancing odorant injection accuracy and the overall system optimization capabilities.

[0132] For example, LSTM predictions are linked to multi-objective optimization. By predicting future flow and pressure trends, the LSTM model provides forward-looking input for multi-objective optimization. For example, if the LSTM predicts a 20% increase in flow over the next 10 minutes, the optimization model can pre-calculate adjustments to valve opening and pump power. This prevents a surge in resource consumption (f2(x)) caused by temporary overshoots while ensuring that the equipment load (f3(x)) remains within limits. Furthermore, the prediction results are combined with real-time data to dynamically adjust the objective function weights. For example, if a sudden pressure drop is predicted, the accuracy target (w1) weight is temporarily increased, forcing injection compensation to prevent concentration runaway.

[0133] In this exemplary embodiment, the objective function can be parameterized when applied. For example, the revised resource consumption target (f2(x)) explicitly includes energy costs (power consumption of pumps and valves) and odorant costs (usage × unit price), which are directly related to parameters such as flow rate and valve opening. For example, for every 10% increase in valve opening, the energy cost increases by 5%, forcing the system to adopt a gradual adjustment when the flow rate fluctuates. The operating intensity coefficient I(x) in the equipment maintenance target (f4(x)) reflects the overload state of the equipment (for example, the coefficient doubles when the valve opening is >80%). The system reduces the injection rate accordingly to extend the life of the equipment, and switches to backup equipment in advance through LSTM prediction.

[0134] Another example is dynamic regulation in a real-world scenario. In a scenario where the primary pressure drops by 10% and the flow rate is predicted to increase by 15%, the system prioritizes safety targets (w1 = 0.4), increases the pressure correction factor to 1.1 times to compensate for the concentration, and pre-starts a backup pump to account for the flow rate change. The optimization results increase injection volume by 12% and energy consumption by 8%, while keeping both equipment load and concentration errors within thresholds. This process demonstrates the synergy between multi-objective optimization and LSTM prediction: predicted data guides weight allocation, while real-time data corrects decision variables, ultimately achieving a global optimization of safety, cost, and equipment health.

[0135] In this application, multi-objective optimization indirectly maps environmental parameters such as flow and pressure to the odorizing dosage adjustment strategy through weight allocation and constraints, while the LSTM prediction model provides temporal trend prediction for this process. The combination of the two enables the system to respond to parameter changes in real time while proactively avoiding conflicting objectives, ultimately achieving a balance between safe concentration, resource efficiency, and equipment lifespan. This integrated design is conducive to improving the intelligent level of odorizing agent control, enabling it to maintain high precision and robustness in complex pipe network environments.

[0136] In this application, before using the LSTM model to predict environmental changes, the LSTM model can be trained. For example, the LSTM model can be trained through supervised learning, with the goal of minimizing the prediction error, that is, the difference between the model's output value and the true value. Figure 2 This is a flow chart of the LSTM model training process according to an embodiment of the present invention. Figure 2 As shown, the training process includes the following steps:

[0137] Step 21: Data Preprocessing and Preparation. First, the model requires historical operational data on pipeline status for training. Raw data (such as flow, pressure, and temperature) is preprocessed, including normalization, filling missing values, and smoothing, to ensure that the data is suitable for LSTM model input.

[0138] Step 22: Data input and target setting. For time series data, the input data is the environmental data for the current moment and several moments before, while the target is the environmental data corresponding to the future moment. For the odorant injection system, the input includes flow rate, pressure, temperature, and other data from the previous moments, while the target is the predicted value of flow rate, pressure, temperature, and other data for a period of time in the future.

[0139] Step 23: Forward propagation. During the forward propagation phase, the input data propagates through multiple layers of the LSTM network to generate a predicted value. Each LSTM unit generates the current output based on the current input, the state at the previous moment, and the cell state.

[0140] Step 24: Error calculation: The error between the predicted result and the true value (target value) is calculated using a loss function, where the loss function used is the mean square error.

[0141] Step 25: Backpropagation. Through the backpropagation algorithm, the error is propagated backward through the network to update the model's weights and parameters. The LSTM model uses a gradient descent algorithm (such as the Adam optimizer) to minimize the error. Backpropagation involves adjusting the weights of all gates (input gate, forget gate, output gate) to better predict future values.

[0142] Step 26: Training iterations. This entire process will iterate multiple times, with each iteration the model will improve its prediction accuracy by updating its weights. During each training session, the network calculates errors and adjusts its parameters to continuously optimize the model's performance.

[0143] Step 27: Model Validation and Adjustment. During the training process, use cross-validation or a validation set to evaluate the model's effectiveness and prevent overfitting. The training process ends when the model's predictive performance meets expectations.

[0144] The trained LSTM model can be used in practical applications to predict future environmental changes. For odorant injection systems, the trained LSTM model can predict future trends based on real-time data (such as flow rate, pressure, and temperature), allowing for dynamic adjustments to the odorant injection rate to ensure it adapts to environmental changes and maintains the natural gas odorant concentration within a safe range.

[0145] In summary, the LSTM model's unique gating mechanism effectively captures long-term dependencies in time series data. Its training process includes data preprocessing, forward propagation, error calculation, backpropagation, and multiple iterations, ultimately generating a model capable of accurately predicting future data. In the odorant injection system, the application of the LSTM model enables real-time prediction of environmental changes, such as flow rate and pressure, and dynamic adjustment of the odorant injection rate, ensuring system safety and stability.

[0146] In some embodiments, the odorant control module is specifically configured to:

[0147] Obtaining the odorant injection amount of each section of the pipeline determined by the intelligent calculation module;

[0148] With the optimization goal of minimizing the error of the odorant injection amount of the entire system, the odorant injection amount of each section of the pipeline is optimized.

[0149] In this exemplary embodiment, based on the aforementioned time series data analysis and LSTM prediction results, the intelligent computing module also applies predictive modeling and error correction methods. After predicting future environmental changes, the system can make advance adjustments based on the predicted data generated by the LSTM model to correct potential injection errors. This method can effectively reduce the slow response caused by sudden environmental changes (such as sudden changes in air pressure or fluctuations in gas composition) and adjust the odorant injection strategy in advance, thus avoiding inaccuracies or safety hazards caused by prediction errors.

[0150] In this exemplary embodiment, the system continuously monitors the deviation between the actual and predicted odorant injection rates during the injection process and optimizes itself based on this feedback. By comparing this deviation with real-time environmental data, the system can promptly detect and correct any deviations. Each adjustment and optimization is fed back to the adaptive learning algorithm via the intelligent computing module, continuously improving the system's overall accuracy and response speed. This continuous optimization process helps ensure that the system maintains optimal performance over the long term and can flexibly adapt to varying operating conditions and environmental changes.

[0151] Through the coordination of the above steps, the system can not only achieve efficient and accurate odorant injection, but also has the ability of adaptive adjustment, dynamic optimization and real-time feedback, and can maintain efficient and stable operation in complex environments.

[0152] In this exemplary embodiment, after the intelligent computing module determines the odorant injection amount, the odorant control module precisely adjusts the odorant injection device. The odorant control module, comprised of precision valves, pumps, and flow controllers, automatically adjusts the odorant injection amount based on the intelligent computing module's instructions. The injection device further fine-tunes the odorant injection amount based on feedback from flow and pressure sensors, ensuring that the natural gas odorant concentration in each pipeline section meets the specified standards. This allows the control device to precisely execute the intelligent computing module's instructions and automatically adjust the odorant injection amount without manual intervention, thereby reducing human error and ensuring accurate odorant addition.

[0153] In this application, after the intelligent computing module receives data from the sensor, it calculates the optimal odorant injection amount based on real-time environmental data (such as natural gas flow, pressure, temperature, etc.) and historical operating data. During the calculation process, the intelligent computing module will comprehensively consider the flow and pressure changes in different pipeline sections based on the set safety standards and target odorant concentration, and determine the odorant dosage for each pipeline section. This decision-making process uses an adaptive control algorithm to dynamically optimize the injection amount to adapt to the ever-changing pipeline conditions and external environmental factors. The end result is a precise injection amount, ensuring the safe and efficient use of natural gas.

[0154] Among them, the odorant injection amount can be adjusted when the injection amount error is taken into account. The process is as follows:

[0155] x represents the set of decision variables that affect the odorant injection amount, including equipment parameters, control parameters, etc.

[0156] C target is the target odorant concentration;

[0157] C actual (i) is the actual odorant concentration in the i-th section of the pipeline;

[0158] Q(i) is the flow rate in the i-th pipe section;

[0159] P(i) is the pressure of the i-th section of the pipeline;

[0160] f(x) represents the amount of odorant injected. The error between the target odorant concentration and the actual concentration is ε(i) = C actual (i)-C target , where ε(i) is the concentration error of the i-th pipeline.

[0161] The optimization goal is to minimize the error in the injection amount of the odorant in the entire system, and to optimize the injection amount of the odorant in each section of the pipeline, specifically including:

[0162] Adjustment of odorant injection amount, including:

[0163] Among them, f inject (i) is the odorant injection rate of the i-th section of the pipeline, k1 is a proportionality coefficient that represents the amount of odorant required per unit flow rate and pressure;

[0164] Considering safety standards and adjustment factors, determine the odorant injection rate f for the i-th section of pipeline adjust (i) is:

[0165] f adjust (i) = f inject (i)·(1+α1·ε(i)); This expression represents the basic injection requirement f inject Based on the error ε(i) between the actual odorant concentration in that section of pipeline and the target concentration, the injection volume is appropriately adjusted to ensure that the minimum safety concentration is met or slightly exceeded. The safety standard refers to the minimum odorant concentration threshold set by the system according to national standards or industry specifications to ensure that the terminal natural gas has sufficient odor recognition. The adjustment factor is a dynamic coefficient used in the system to compensate for odorant concentration deviations. Its value is determined based on the error between the actual odorant concentration in each section and the target concentration, and is used to dynamically adjust the injection volume in each section to meet the safety standard.

[0166] Where α1 is an adjustment factor that represents the effect of concentration error on injection volume. It is a sensitivity coefficient or adjustment factor that represents the response strength of injection volume to error. If the error is large, more odorant dosage needs to be adjusted to ensure a safe concentration.

[0167] The final odorant injection amount determination formula can be expressed as:

[0168] f final (i) = f adjust (i)+f external (i); where f external (i) The impact of external environmental factors on the injection volume (e.g. temperature changes, external climate, etc.);

[0169] In this example, dynamic optimization is used to minimize the error in the odorant injection rate of the entire system in each pipeline segment while adapting to changes in flow and pressure. Therefore, the overall goal can be expressed as:

[0170] Where γ is the weight coefficient between the adjustment of odorant injection amount and error, and N is the total number of pipeline segments.

[0171] In this embodiment, the objectives defined in the multi-objective optimization, such as f1(x) (accuracy) and f2(x) (resource consumption), indirectly control the injection volume by influencing the decision variable x (e.g., the proportional coefficient k1, the adjustment factor α1, and the weight γ). For example, if the optimization results require reducing resource consumption (f2(x) has a high weight), the system will dynamically reduce the γ value to reduce the injection volume while allowing for slight concentration errors. Conversely, if safety priority is high (f1(x) has a high weight), α is increased to strengthen error compensation to ensure that the concentration meets the target.

[0172] In this embodiment, the LSTM model provides forward-looking input for injection calculation by predicting the future trends of parameters such as flow rate and pressure. For example, if the LSTM predicts that the flow rate will increase by 20% in the next 5 minutes, the system will adjust f in advance according to the increase in Q(i). inject (i) rather than passively waiting for real-time data feedback. This predictive capability makes the f in the injection formula external (i) (external environmental correction term) can incorporate future environmental changes, such as pre-increasing the injection volume before a temperature drop to offset the risk of increased concentration caused by gas contraction, thereby reducing the pressure of subsequent error correction.

[0173] The concentration error ε(i) is defined as the deviation between the real-time measurement value and the target value, and the LSTM prediction is used to proactively correct possible errors in the future. For example, if the LSTM predicts that the pressure is about to drop, the system will increase f in the current injection calculation in advance. adjust (i) compensation, so as to avoid the actual concentration Cactual (i) Deviation from the target value. This mechanism incorporates the forecast error (future deviation) into the current decision, so that ε(i) not only reflects the historical error but also pre-controls the future state through the forecast data, thus achieving closed-loop optimization.

[0174] Among them, the dynamic optimization goal It is a simplified practical version of the multi-objective function F(x). Among them, ε(i) directly corresponds to the accuracy target of f1(x), γ·f final (i) implicitly includes resource consumption (f2(x)) and device load (f3(x)). For example, the value of γ is determined by the weights w2 and w3 in the multi-objective optimization. This layered design allows real-time control (step 4) and global optimization (step 3) to be both independent and collaborative: real-time control rapidly responds to environmental changes, while global optimization periodically adjusts parameters to adapt to long-term goals.

[0175] Injection rate adjustment is the result of the combined efforts of objective function optimization and LSTM prediction. The multi-objective function defines the control direction (e.g., safety or cost priority) through weight assignment, while the LSTM prediction provides input for future environmental parameters. Together, these two drive the dynamic adjustment of the aforementioned formula parameters (e.g., k1, α1, and γ). This collaborative mechanism enables the system to not only correct current concentration errors in real time but also proactively mitigate forecast deviations, ultimately achieving global optimization across multiple dimensions, including safety, efficiency, and equipment lifespan.

[0176] After the intelligent calculation module determines the odorant injection amount, it transmits this decision to the odorant control module via the data communication module. Upon receiving the command, the odorant control module activates precise valves, pumps, and flow controllers, adjusting the injection mechanism accordingly. This process is automated by built-in electric valves, precision pumps, and high-precision flow controllers, ensuring that the calculated amount of odorant is accurately injected into the system. This process requires no human intervention and is fully controlled by the intelligent system, minimizing human error and operational errors.

[0177] During the odorant injection process, the odorant injection device not only adjusts the injection volume according to the instructions of the intelligent computing module but also monitors the flow and pressure in the pipeline in real time. Through the installed flow and pressure sensors, the system can obtain actual pipeline operating data in real time. If the sensors detect any deviation or anomaly (such as unstable flow or pressure fluctuations outside the preset range), the system will immediately provide feedback and adjust the odorant injection volume to ensure that the natural gas odorant concentration in each section of the pipeline always meets the set standards. This real-time adjustment can cope with rapid changes in the environment and subtle fluctuations in the equipment, maintaining accurate odorant delivery.

[0178] During the odorant injection process, all operational data, including injection volume, flow rate, and pressure, is recorded and stored. This data is not only used in real-time to adjust and optimize injection strategies, but also for subsequent operational analysis and troubleshooting. Through the system's intelligent algorithms and data storage capabilities, all operation histories and adjustment records are traceable, ensuring transparency and controllability throughout the entire process. At any time, operators or relevant regulatory agencies can review odorant injection records to verify compliance with safety and legal requirements.

[0179] The odorant control module is also equipped with a self-check function, which monitors the device's status in real time during the injection process. If sensors detect abnormal data or a device malfunction, the system automatically activates a fault warning function, notifying the operator to conduct inspection or repairs. Furthermore, the system uses intelligent algorithms to analyze potential fault points and take preventive measures before a failure occurs, preventing damage to the device or improper odorant application. This early warning mechanism improves operational safety and stability, ensuring the continuity and accuracy of the odorant injection process.

[0180] Through these steps, the system can accurately, efficiently, and stably control odorant injection. Furthermore, the system can adjust based on real-time data to ensure that the odorant concentration in natural gas meets safety standards. It also provides data traceability, fault warnings, and real-time monitoring, comprehensively improving the system's intelligence and operational efficiency.

[0181] When a sensor or device in the system malfunctions, the system automatically detects it through its built-in fault diagnosis module. By analyzing the abnormal data returned by the sensors, the module identifies the device fault and generates a fault warning. At this point, the remote control module notifies maintenance personnel to intervene, or the system automatically adjusts parameters to maintain normal odorant injection. The advantage of this step is that the system intelligently adjusts operations based on the fault diagnosis results, ensuring that the system remains as functional as possible even in abnormal conditions, without the need for human intervention.

[0182] in, Figure 3 This is a flow chart of fault diagnosis and automatic adjustment according to an embodiment of the present invention. Figure 3 As shown in FIG, fault diagnosis and automatic adjustment specifically include the following steps:

[0183] Step 31, the fault diagnosis module is started. During the operation of the system, when any sensor or device fails, the system will automatically start the fault diagnosis module. The fault diagnosis module will monitor the working status of the sensors and equipment in real time, and detect the data fed back by the sensors. When the data fed back by the sensor exceeds the preset normal range, or there is a data abnormality (such as exceeding the standard error, signal loss, etc.), the fault diagnosis module will immediately mark the abnormality and generate a fault alarm. This process ensures that the fault can be identified at an early stage and can trigger the subsequent diagnosis and repair process in time. In this way, the fault diagnosis module automatically detects anomalies based on real-time data, can identify potential faults in advance and generate alarms, avoiding manual inspections and delayed responses.

[0184] Step 32, fault type and location identification. When the fault diagnosis module finds data anomalies, the system will analyze the source of the problem and further identify the specific type of fault and the location of the occurrence. For example, the system will determine whether the fault comes from the sensor itself (sensor failure, data error, etc.) or other equipment (valve or pump failure of the odorant injection device). By comparing with historical data, the system can accurately identify the fault point and distinguish different types of faults. The identification of fault type and location is the basis for subsequent automatic adjustment or manual intervention. In this way, through intelligent analysis, the system can not only detect faults, but also accurately identify the fault type and location, reduce misdiagnosis and improve repair efficiency.

[0185] Step 33, automatic adjustment and compensation. In certain fault conditions, the system can maintain normal operation by automatically adjusting parameters. For example, when a sensor fails or some data deviations occur, the system will automatically adjust the odorant injection volume based on the backup sensor data or the historical operating data pattern of the pipeline operating status to ensure that the system continues to operate stably in the short term. In addition, the system can also adjust the operating status of the equipment, such as by adjusting the pump flow rate, the odorant injection speed and other parameters to compensate for the impact of the faulty equipment and maintain the overall performance of the system. Thus, through the adaptive automatic adjustment capability, the system can adjust the operating parameters according to the real-time fault diagnosis results without human intervention, ensuring that the system continues to operate in the event of partial equipment failure and maintaining the accuracy of the odorant injection volume.

[0186] Step 34, fault warning and remote intervention. After the system identifies the fault and makes preliminary adjustments, the remote control module will transmit the fault information to the maintenance personnel and notify them to intervene. At this time, the system will provide detailed fault information, diagnostic results and the current automatic adjustment status to help maintenance personnel quickly locate the problem and repair it. If the fault is more complex or has a large impact, the remote control module can start the emergency treatment procedure to ensure the safety and stability of the system. In this embodiment, the emergency treatment procedure refers to a set of preset emergency response mechanisms automatically triggered by the remote control module when the system detects a major fault, an abnormality in key equipment, or the odorant injection system fails to respond normally. The procedure may include but is not limited to: immediately suspending the operation of related equipment, switching to a safe standby mode, enabling redundant odorizing devices, reducing the system operating load, limiting the odorant injection rate, sending alarm information to relevant responsible persons and supervision platforms, etc., to avoid safety hazards or environmental risks brought about by the continuous operation of the system. At the same time, the emergency handling program can also link with the operation and maintenance interface on the cloud platform, retrieve historical data for rapid fault location, and initiate local or remote emergency control strategies to ensure that the system still has basic operational guarantee capabilities under abnormal conditions, thereby maximizing the safety and stability of the natural gas transmission and distribution process.

[0187] Step 35: Data Recording and Traceability. The fault diagnosis module records all fault events and their diagnostic process, including fault type, location, corrective measures, and recovery status. All data is uploaded to the cloud platform in real time and encrypted using blockchain technology to ensure data immutability and provide a complete traceability record. Every fault adjustment and repair in the system can be traced back to the specific time, operation, and response measures to ensure operational transparency and compliance. Blockchain technology enables tamper-proof storage of fault events, ensuring the integrity and transparency of fault data and facilitating supervision and subsequent review.

[0188] Through the above steps, the system can intelligently diagnose and automatically adjust when a fault occurs, achieving rapid response through remote control. These measures ensure that the system can maintain normal operation to the greatest extent possible even in the event of a fault, avoiding large-scale downtime or safety hazards. Furthermore, the system's data logging and blockchain storage ensure operational transparency and compliance, providing a reliable basis for subsequent maintenance.

[0189] In this embodiment, the remote control module allows operators to monitor the operating status of the entire system in real time. This module supports viewing sensor data, odorant injection levels, device status, and other information via a cloud platform or dedicated terminal, and allows remote adjustment of system operating parameters. Operators can also use the remote platform to optimize and adjust odorant injection as necessary. The remote diagnostic module also provides troubleshooting advice for the device, reducing the need for manual intervention.

[0190] in, Figure 4 This is a flow chart of remote monitoring and management according to an embodiment of the present invention. Figure 4 As shown in the figure, remote monitoring and management specifically includes the following steps:

[0191] Step 41, real-time data collection and transmission. The system transmits sensor data, odorant injection volume, equipment status and other information to the cloud platform or dedicated terminal in real time through the remote control module. The data collected by all sensors (such as temperature, pressure, flow and odorant concentration sensors) will be encrypted and transmitted to ensure the security and confidentiality of the data. Through this data transmission mechanism, operators can view the real-time data of each sensor at any time and understand the operating status of the system in real time, including the flow and odorant concentration of each pipeline section. This sub-step ensures uninterrupted data flow throughout the system and can achieve real-time monitoring on a global scale. In this way, encrypted transmission technology is used to ensure the security and integrity of the data, while ensuring that operators can obtain system data in real time without on-site operation, thereby improving the convenience and reliability of remote management.

[0192] Step 42, remote adjustment and optimization. Through the cloud platform or dedicated terminal, the operator can view the system operation status in real time and make remote adjustments and optimizations as needed. The operator can not only view the current odorant injection volume, equipment status, sensor data, etc., but also directly make necessary adjustments to the system. For example, the operator can adjust the odorant injection volume based on the monitoring results, and optimize parameters such as pipeline flow or pressure to ensure that the system operates in the optimal state. In addition, the system will provide optimization suggestions based on real-time data and preset optimization goals using intelligent optimization algorithms (such as machine learning-based prediction models or multi-objective optimization algorithms). These optimization suggestions include adjusting the odorant injection volume, optimizing flow, adjusting equipment load, and other operations to ensure that the system operates at the highest efficiency. The operator can choose whether to implement these optimization suggestions, or modify the optimization strategy according to specific needs. The system not only provides remote viewing functions, but also allows operators to make remote adjustments and optimizations. Through the automatic optimization function, the system can automatically improve the odorant injection process without intervention, reducing the risk of human operation.

[0193] Step 43: Remote diagnosis and troubleshooting suggestions. The remote diagnosis module continuously monitors the status of equipment and sensors, promptly identifies faults, and generates diagnostic reports. When the system detects a device or sensor fault, the remote diagnosis module automatically analyzes the cause and provides treatment suggestions. The fault diagnosis report and treatment plan will be sent to the operator via the cloud platform or dedicated terminal to ensure that the problem can be resolved promptly. Through this process, operators can quickly understand the fault situation and take appropriate measures in a remote environment, reducing the frequency and time of on-site maintenance. Through the intelligent remote diagnosis module, the system can automatically analyze faults and provide treatment suggestions. Operators can not only view the status of the equipment remotely, but also take appropriate measures based on the diagnostic results, thereby reducing the need for on-site operations and response time.

[0194] Specifically, the remote diagnosis module automatically analyzes the cause of the fault and provides treatment suggestions as follows:

[0195] The core function of the remote diagnostic module is to identify faults, automatically analyze the causes of faults, and ultimately generate treatment recommendations through real-time monitoring and analysis of device and sensor data. There are many ways to implement this function, including the following:

[0196] Rule-based expert systems use a set of pre-set expert rules and thresholds. When the output data from a device or sensor exceeds a set safety range, the system triggers the rules, analyzes the cause of the failure, and provides treatment recommendations. For example, if a sensor reading exceeds the safety range, the system will infer the possible fault type based on empirical rules, such as device overload or sensor failure, and provide relevant treatment options (such as restarting the device or adjusting the device load). This approach has the advantage of simplicity and directness, but it relies on manually formulated rules and lacks flexibility and automatic learning capabilities.

[0197] Machine learning-based fault detection, in which the remote diagnostic module learns normal and fault modes from historical operating data using machine learning algorithms (such as classification algorithms, clustering algorithms, etc.). By training the model, the system can predict the type of fault of the device or sensor based on the sensor data and device status collected in real time, and provide treatment suggestions. For example, the system learns the difference between normal behavior and fault behavior in device operation by training the model using supervised learning technology. When real-time data deviates from the known patterns in the training data, the system will identify it as a fault and provide a diagnostic result. For example, when the power curve of a device does not match the historical pattern, the system determines that the device is overloaded and recommends that the operator check the device load. This method can provide a high level of automation and can continuously improve diagnostic accuracy as data accumulates.

[0198] Data-driven anomaly detection and diagnosis: This approach uses data-driven algorithms, such as anomaly detection algorithms (e.g., isolation forests, clustering algorithms, etc.). This approach models normal behavior by modeling historical data from devices and sensors. The system monitors device operating data in real time and compares it with known normal patterns, discovering abnormal patterns in the data and identifying faults. For example, when fluctuations in parameters such as device temperature and pressure exceed historical fluctuation ranges, the system can automatically identify the anomaly and infer the possible fault type, such as device damage or system overload. The system can then generate appropriate action suggestions, such as adjusting the device's operating status or performing equipment maintenance.

[0199] Deep learning-based prediction and diagnosis. Deep learning methods (such as LSTM, convolutional neural networks (CNN), etc.) can be used to automatically extract features from large amounts of historical data and perform multi-dimensional fault prediction and diagnosis. By training deep neural networks, the system can identify complex patterns and long-term dependencies in the data, thereby predicting possible equipment failures and providing treatment suggestions. For example, the LSTM model is used to predict the status of the equipment over a period of time in the future, and based on the prediction results, it is determined whether there is a potential fault. If the predicted equipment load or temperature trend changes abnormally, the system will automatically infer that the equipment is about to fail and provide treatment suggestions to the operator (such as reducing the load, stopping operation, etc.). The advantage of this method is that it can process highly complex time series data and provide more accurate and comprehensive fault diagnosis.

[0200] Therefore, the remote diagnosis module's automatic analysis of fault causes and provision of treatment recommendations can be implemented through a variety of technologies, including rule-based expert systems, machine learning models, data-driven anomaly detection algorithms, and deep learning models. By combining these multiple approaches, the system's flexibility and intelligence can be enhanced, making fault diagnosis and treatment recommendations more accurate and reliable, thereby reducing the frequency and time of on-site repairs and improving the efficiency of remote management.

[0201] Step 44: Data Recording and Historical Analysis. All remote operations and adjustments will be recorded and stored, forming a complete operation history. This historical data not only facilitates subsequent fault analysis, system tuning, and technical support, but also provides data support for future optimization. Through big data analysis, the system can predict trends based on historical data and provide future optimization solutions. Historical data will be stored on a cloud platform or distributed storage system to ensure long-term data availability and support system upgrades. This comprehensive recording of historical data and big data analysis capabilities will enhance the system's traceability and predictive capabilities, helping operators to achieve more efficient management and optimization in the future.

[0202] Step 45: System Security and Compliance Assurance. To ensure the security and compliance of the remote monitoring system, all remote operations and diagnostic processes are encrypted and utilize multi-factor authentication to ensure only authorized personnel can operate the system. The system also undergoes regular security reviews and compliance checks based on industry standards and regulatory requirements to ensure compliance with relevant security, privacy, and operational specifications. This process enhances the system's protection capabilities and ensures the security of data and operations. By employing multi-layered security measures, such as data encryption and multi-factor authentication, the system ensures the security of remote monitoring and management, protecting the system from external threats and improper operation.

[0203] Through these steps, the system not only provides real-time monitoring and adjustment capabilities in a remote environment, but also improves operational efficiency and flexibility through intelligent remote diagnosis and automatic optimization. Furthermore, through historical data recording and big data analysis, the system can ensure long-term stable operation and anticipate and address potential risks.

[0204] The system regularly uploads all operational data, sensor feedback, fault records, and other information to a cloud server for storage. This data is encrypted and stored, and historical data can be queried and analyzed at any time. By analyzing historical data, the intelligent computing module continuously optimizes odorant injection strategies for long-term performance improvement and system tuning. Furthermore, the system's historical data provides a valuable basis for future fault diagnosis, equipment maintenance, and optimization. The advantage lies in combining historical data with intelligent analysis algorithms, enabling the system to self-learn and optimize over the long term, thereby continuously improving performance.

[0205] The system's flexible modular design and open interfaces allow for expansion and upgrades based on future needs. For example, new sensors, control modules, or other devices can be seamlessly integrated into the system without impacting the stable operation of the existing system. Furthermore, as technology evolves, the system can incorporate new intelligent algorithms and control strategies through software upgrades, maintaining technological advancement.

[0206] Through these steps, the distributed network-based natural gas odorant intelligent control system of the present invention can achieve fully automated and highly intelligent natural gas odorant control, thereby improving the operating efficiency, reliability and safety of the system.

[0207] The present application provides a natural gas odorant intelligent control method based on a distributed network. Figure 5 This is a flow chart of the distributed network-based intelligent control method for natural gas odorant according to an embodiment of the present invention. Figure 5 As shown, it includes:

[0208] Step 51: Collect environmental data of natural gas flowing through the pipeline; wherein the environmental data at least includes pipeline temperature, pipeline pressure, natural gas flow rate, ambient humidity, gas composition in the pipeline, and odorant concentration;

[0209] Step 52: Determine the amount of odorant injected into the pipeline based on the odorant adjustment algorithm in combination with the environmental data and historical operation data of the pipeline operation status;

[0210] Step 53: Adjust the concentration of the odorant in the pipeline according to the amount of the odorant injected into the pipeline.

[0211] The distributed network-based natural gas odorant intelligent control method of the present application can refer to the distributed network-based natural gas odorant intelligent control system described above.

[0212] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0213] See also Figure 6 , an embodiment of the present application further provides an electronic device 600, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the distributed network-based natural gas odorant intelligent control method in the aforementioned method embodiment.

[0214] An embodiment of the present application further provides a non-transitory computer-readable storage medium storing computer instructions for causing the computer to execute the distributed network-based natural gas odorant intelligent control method in the aforementioned method embodiment.

[0215] An embodiment of the present application also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer executes the distributed network-based natural gas odorant intelligent control method in the aforementioned method embodiment.

[0216] Reference below Figure 6 , which shows a schematic structural diagram of an electronic device 600 suitable for implementing the embodiments of the present application. The electronic device 600 in the embodiments of the present application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 6 The electronic device 600 shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0217] like Figure 6 As shown, the electronic device 60 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. Various programs and data required for the operation of the electronic device 600 are also stored in the RAM 603. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0218] The following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touchpad, a key, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. Although the figure shows the electronic device 600 with various devices, it should be understood that it is not required to implement or have all of the devices shown. More or fewer devices may be implemented or have alternatively.

[0219] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network via the communication device 609, or installed from the storage device 608, or installed from the ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the method of the embodiment of the present application are performed.

[0220] The above description is merely a specific embodiment of the present disclosure, but the scope of protection of the present disclosure is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this disclosure should be included in the scope of protection of the present disclosure. Therefore, the scope of protection of the present disclosure should be based on the scope of protection of the claims.

Claims

1. A natural gas odorant intelligent control system based on a distributed network, characterized in that: include: A sensor monitoring module is used to collect environmental data of natural gas flowing through the pipeline; wherein the environmental data at least includes pipeline temperature, pipeline pressure, natural gas flow rate, ambient humidity, gas composition in the pipeline, and odorant concentration; an intelligent computing module configured to receive environmental data transmitted by the sensor monitoring module via a distributed network, calculate an odorant injection control instruction based on an odorant adjustment algorithm in combination with the environmental data and historical operating data of the pipeline operation status, and transmit the odorant injection control instruction to the odorant control module; wherein the odorant injection control instruction includes at least an odorant injection amount in the pipeline that needs to be dynamically adjusted; an odorant control module, configured to adjust the odorant injection amount in the pipeline according to the odorant injection control instruction, so as to adjust the odorant concentration in the pipeline; A remote control module is used to monitor the operating status of the sensor monitoring module, the intelligent computing module and the odorant control module.

2. The distributed network-based natural gas odorant intelligent control system according to claim 1 is characterized in that: The odorant adjustment algorithm includes: Determining a basic odorant injection amount based on the natural gas flow rate, the target odorant concentration, the actual odorant concentration, and the system automatic increase proportional coefficient; determining a pressure correction factor based on the pipeline pressure and the standard pressure; determining a pressure-corrected injection amount based on the pressure correction factor and the basic injection amount of the odorant; determining a temperature correction factor based on the pipeline temperature and a standard temperature; determining a temperature-corrected injection amount based on the temperature correction factor and the pressure-corrected injection amount; determining a composition correction factor based on a methane concentration in the gas composition in the pipeline and a baseline methane concentration; determining a composition-corrected injection amount based on the composition correction factor and the temperature-corrected injection amount; determining a humidity correction factor based on the ambient humidity, saturated humidity, and reference humidity; Based on the humidity correction factor and the component correction injection amount, the odorant injection amount in the pipeline is determined; wherein the standard pressure, the standard temperature, the baseline methane concentration, the saturated humidity and the reference humidity are all the historical operating data.

3. The distributed network-based natural gas odorant intelligent control system according to claim 2 is characterized in that: The odorant control module is specifically used to: Determine the objective function of odorant injection accuracy, resource consumption, equipment load, equipment maintenance cost, environmental friendliness, response speed optimization, and system stability; Constructing a comprehensive objective function based on the odorant injection accuracy objective function, resource consumption objective function, equipment load objective function, equipment maintenance cost objective function, environmental friendliness objective function, response speed optimization objective function, and system stability objective function; Based on minimizing the comprehensive objective function as the optimization goal, the injection pattern of the odorant in the pipeline is used as the function optimization variable to determine the optimized injection pattern of the odorant; Based on the optimized odorant injection pattern, the odorant injection amount in the pipeline is adjusted to adjust the odorant concentration in the pipeline.

4. The distributed network-based natural gas odorant intelligent control system according to claim 3 is characterized in that: The odorant control module is specifically used to: The LSTM original model is trained based on the historical operation data of the pipeline operation status to obtain an LSTM prediction model; Predicting future pipeline natural gas flow and pipeline pressure trends based on the LSTM prediction model; Based on the optimized odorant injection pattern and in combination with the future natural gas flow and pipeline pressure trends of the pipeline, the odorant injection amount in the pipeline is adjusted to adjust the odorant concentration in the pipeline.

5. The distributed network-based natural gas odorant intelligent control system according to claim 2 is characterized in that: The odorant control module is specifically used to: Obtaining the odorant injection amount of each section of the pipeline determined by the intelligent calculation module; With the optimization goal of minimizing the error of the odorant injection amount of the entire system, the odorant injection amount of each section of the pipeline is optimized.

6. The distributed network-based natural gas odorant intelligent control system according to claim 2, characterized in that: The method of determining the basic odorant injection amount based on the natural gas flow rate, the target odorant concentration, the actual odorant concentration, and the system automatic increase proportional coefficient includes: Among them, f base is the basic injection amount of the odorant; C target is the target odorant concentration, C actual is the actual odorant concentration, Q is the natural gas flow rate, and k1 is the system's automatic increase coefficient.

7. The distributed network-based natural gas odorant intelligent control system according to claim 6, characterized in that: Determining the pressure correction factor based on the pipeline pressure and the standard pressure includes: Where β(P) is the pressure correction factor, P is the pipeline pressure, and P0 is the standard pressure; The step of determining the pressure-corrected injection amount based on the pressure correction factor and the basic injection amount of the odorant includes: f pressure =f base β(P); where f pressure is the pressure-corrected injection volume, f base This is the base injection amount of the odorant.

8. The distributed network-based natural gas odorant intelligent control system according to claim 7, characterized in that: The determining of the temperature correction factor based on the pipeline temperature and the standard temperature includes: Where γ(T) is the temperature correction factor, T0 is the standard temperature, and T is the pipe temperature; The determining of the temperature correction injection amount based on the temperature correction factor and the pressure correction injection amount includes: f temp =f pressure γ(T); where f temp is the temperature correction injection amount, f pressure Correct the injection volume for pressure.

9. The distributed network-based natural gas odorant intelligent control system according to claim 8, characterized in that: The determining of the composition correction factor based on the methane concentration in the gas component in the pipeline and the reference methane concentration includes: Among them, δ(C CH4 ) is the composition correction factor, C ref is the baseline methane concentration, C CH4 is the methane concentration in the gas composition in the pipeline, and α is the proportional adjustment coefficient; The determining of the component correction injection amount based on the component correction factor and the temperature correction injection amount includes: f gas =f temp ·δ(C CH4 ); where f gas is the component correction injection amount, f temp Correct the injection volume for temperature.

10. The distributed network-based natural gas odorant intelligent control system according to claim 9, characterized in that: The determining of the humidity correction factor based on the ambient humidity, the saturated humidity, and the reference humidity includes: Where η(H) is the humidity correction factor, H sat is the saturated humidity, H is the ambient humidity, H ref is the reference humidity, k H is an adjustable proportional coefficient; The step of determining the injection amount of the odorant in the pipeline based on the humidity correction factor and the component correction injection amount includes: f humidity =f gas ·η(H); where f humidity is the amount of odorant injected into the pipeline, f gas Correct injection volume for the component.

Images