Compressor control management system based on Internet of Things

Abstract

The invention discloses a compressor control management system based on the Internet of Things, and relates to the technical field of industrial automation and intelligent control. The compressor control management system based on the Internet of Things comprises a multi-dimensional data acquisition module which acquires multi-dimensional data and historical data of a compressor in real time through an Internet of Things technology multi-dimensional sensor. The suction valve model building unit builds the relation between the suction valve opening degree and the compressor load, energy efficiency and suction flow based on the long short-term memory network, and the exhaust valve model building unit builds the relation between the exhaust valve opening degree and the exhaust pressure, flow and compressor energy efficiency based on support vector regression. The control module dynamically adjusts the opening degree of an air suction valve through a particle swarm optimization algorithm, adjusts the opening degree of an exhaust valve through a fuzzy control algorithm, and optimizes the energy efficiency and stability of the compressor. Automatic adjustment of the opening degree of the compressor is achieved, the energy efficiency is improved, the energy consumption is reduced, and the automation level and the self-adaptive capacity of the system are enhanced.

Description

Technical Field

[0001] This invention relates to the field of industrial automation and intelligent control technology, specifically to a compressor control and management system based on the Internet of Things. Background Technology

[0002] With the accelerating pace of industrialization, energy consumption and environmental protection have become global concerns. Compressors, as crucial equipment in industrial production, play a key role in many fields, including air conditioning, refrigeration, air compression, and petrochemicals. The efficient operation of compressors is of great significance for energy conservation, emission reduction, and improving production efficiency. Meanwhile, with the rapid development of IoT technology, the Industrial Internet of Things (IIoT) provides new opportunities for the intelligent management and optimization of compressors. By combining compressors with multi-dimensional sensors, data processing platforms, and intelligent control systems, real-time monitoring, precise control, and efficient management can be achieved, thereby reducing energy consumption, extending equipment lifespan, and improving production stability, driving the development of industrial equipment towards intelligence and green practices.

[0003] Currently, traditional compressor control and management methods mostly rely on manual adjustments or fixed algorithms, which suffer from drawbacks such as poor real-time response and low optimization levels. Traditional control systems struggle to make rapid and precise adjustments based on compressor load fluctuations, discharge pressure, and energy efficiency requirements, resulting in low compressor operating efficiency and high energy consumption. Furthermore, the lack of data-driven predictive models hinders the intelligent and personalized prediction of equipment failures and the development of maintenance plans. Even with automated control, comprehensive optimization and dynamic adaptation to complex operating conditions are lacking, failing to achieve optimal compressor operating efficiency. Therefore, existing methods cannot fully realize the energy-saving potential of compressors and improve their operational stability, necessitating an intelligent, data-driven control and management system. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides an Internet of Things-based compressor control and management system, which solves the problems mentioned in the background.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an IoT-based compressor control and management system, comprising the following modules: a multi-dimensional data acquisition module, a model building module, and a control module; the modules are used to connect to multi-dimensional sensors via the IoT to acquire multi-dimensional data and historical multi-dimensional data of the compressor; the model building module includes an intake valve model building unit and an exhaust valve model building unit; the intake valve model building unit is used to establish a mathematical relationship between the intake valve opening degree and the compressor load, energy efficiency, and intake flow rate through a long short-term memory network based on the multi-dimensional data and historical multi-dimensional data of the compressor, and to construct an intake valve model for predicting the impact of intake valve opening degree adjustment on compressor efficiency under different loads; the exhaust valve model building unit is used to... Multi-dimensional and historical multi-dimensional data of the compressor are used to establish a mathematical relationship between the discharge valve opening and the compressor discharge pressure, discharge flow rate, and compressor energy efficiency through support vector regression. This constructs a discharge valve model to predict the impact of discharge valve opening adjustment on compressor energy efficiency and stability under different load and discharge pressure conditions. The control module includes a suction valve control unit and a discharge valve control unit. The suction valve control unit dynamically adjusts the suction valve opening based on the suction valve model, combined with real-time load data and compressor energy efficiency requirements, using a particle swarm optimization algorithm. The discharge valve control unit dynamically adjusts the discharge valve opening based on a prediction model generated by the discharge valve model construction unit, combined with real-time discharge pressure, discharge flow rate, and compressor load data, using a fuzzy control algorithm.

[0006] Furthermore, the multi-dimensional data includes, but is not limited to, compressor load, exhaust temperature, exhaust pressure, intake volume, intake and exhaust valve opening, and motor speed.

[0007] Furthermore, the specific process of establishing the mathematical relationship between the intake valve opening and compressor load, energy efficiency, and intake flow rate through a Long Short-Term Memory (LSTM) network is as follows: Preprocessing is performed based on multi-dimensional compressor data and historical multi-dimensional data, including data cleaning, data normalization, and feature extraction, converting the data into a time series format. Historical data is divided into multiple samples according to a set time window, with each window serving as input data for the LTM network. Input data at each time step is segmented through windows, containing feature data on compressor load, energy efficiency, and intake flow rate. The network structure of the LTM model is set, including setting the input layer to accept feature data such as compressor load, energy efficiency, and intake flow rate; setting the number of network layers and the number of neurons in each layer to capture temporal dependencies between data; and setting the output layer to predict the intake valve opening. The LTM network model is trained using historical multi-dimensional data, including setting a loss function to measure the difference between predicted and actual values ​​and updating the model weights through backpropagation and gradient descent algorithms.

[0008] Furthermore, the specific process of constructing the intake valve model is as follows: through a trained long short-term memory network model, the mathematical relationship between the intake valve opening degree and the compressor load, energy efficiency, and intake flow rate is obtained; based on the obtained mathematical relationship, combined with the real-time collected compressor load, energy efficiency, and intake flow rate, the intake valve opening degree under different conditions is calculated, and an intake valve model between intake valve opening degree adjustment and compressor efficiency is generated.

[0009] Furthermore, the specific process of establishing the mathematical relationship between the exhaust valve opening and the compressor exhaust pressure, exhaust flow rate, and compressor energy efficiency through support vector regression is as follows: Based on historical multi-dimensional data, the historical data is preprocessed using a support vector regression model, including data cleaning, normalization, and feature extraction; the training data is fitted using the support vector regression model to establish the mathematical relationship between the exhaust valve opening and the compressor exhaust pressure, exhaust flow rate, and energy efficiency; and the accuracy and generalization ability of the model are ensured by optimizing the hyperparameters in the SVR model, including the penalty factor and kernel function type.

[0010] Furthermore, the specific process of constructing the exhaust valve model is as follows: Using a trained support vector regression model, the mathematical relationship between the exhaust valve opening and the compressor's exhaust pressure, exhaust flow rate, and energy efficiency is obtained; based on the obtained mathematical relationship, combined with real-time collected data such as compressor exhaust pressure, exhaust flow rate, and energy efficiency, the exhaust valve opening under different operating conditions is calculated; an exhaust valve model is generated that relates the exhaust valve opening adjustment to the compressor's energy efficiency and stability, used to dynamically adjust the compressor's exhaust valve opening and optimize the compressor's operating efficiency and stability.

[0011] Furthermore, the specific process of dynamically adjusting the suction valve opening using the particle swarm optimization algorithm is as follows: Real-time compressor load, energy efficiency, and suction flow data are acquired and input into the suction valve model to calculate the current suction valve opening; based on the real-time data and compressor energy efficiency requirements, a target suction valve opening is set, and the difference between the target opening and the current opening is calculated; the particle swarm in the particle swarm optimization algorithm is initialized, and a fitness function is set for each particle, which measures how close the suction valve opening adjusted by the particle is to the target opening; the particle positions are iteratively updated using the particle swarm optimization algorithm, and the fitness function is calculated in each iteration, selecting the particle with the best fitness as the optimal solution; when the iterative process of the particle swarm converges, the optimal solution obtained is the optimized suction valve opening, used to dynamically adjust the suction valve and optimize the compressor's operating efficiency.

[0012] Furthermore, the specific process of dynamically adjusting the exhaust valve opening using a fuzzy control algorithm is as follows: Real-time collected exhaust pressure, exhaust flow rate, and compressor load data are acquired as input variables for the fuzzy control algorithm; based on the actual data of exhaust pressure, exhaust flow rate, and compressor load, the input variables are transformed into fuzzy sets through fuzzification processing; a fuzzy control rule base is established, and based on the exhaust valve model, the relationship between exhaust pressure, exhaust flow rate, and compressor load is determined, thus determining the direction and extent of exhaust valve opening adjustment; fuzzy inference is used to reason about the fuzzy control rules, calculating the adaptive output value of the exhaust valve, i.e., the exhaust valve opening adjustment amount; the fuzzy inference result is defuzzified to obtain the actual exhaust valve opening adjustment value, and the exhaust valve opening is dynamically adjusted through the control system to optimize the compressor's energy efficiency and stability.

[0013] The present invention has the following beneficial effects: (1) The compressor control and management system based on the Internet of Things (IoT) uses a model building module with suction valve model building units and discharge valve model building units to establish accurate mathematical relationship models based on multi-dimensional and historical data of the compressor, and to make dynamic predictions for the suction valve and discharge valve respectively. The suction valve model can predict the impact of suction valve opening adjustment on compressor efficiency under different loads, while the discharge valve model can predict the impact of discharge valve opening adjustment on compressor energy efficiency and stability. Through the establishment of these models, accurate data support and optimization basis are provided for the compressor control system, ensuring that the system can make precise adjustments according to the real-time status, thereby improving overall energy efficiency and stability.

[0014] (2) The IoT-based compressor control and management system, through the intake valve control unit and the exhaust valve control unit, can perform fine adjustment based on the established intake and exhaust valve models and real-time data, using particle swarm optimization algorithm and fuzzy control algorithm. The intake valve control unit dynamically adjusts the intake valve opening through optimization algorithm to ensure the best match between compressor load and energy efficiency, significantly improving the system's operating efficiency and reducing energy consumption; while the exhaust valve control unit uses fuzzy control algorithm to adjust the exhaust valve opening, which can accurately respond to compressor load fluctuations and exhaust pressure changes, enhance system stability, reduce the risk of failure, and optimize equipment operation management.

[0015] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0016] Figure 1 This is a flowchart of the compressor control and management system based on the Internet of Things of this invention. Detailed Implementation

[0017] This application's embodiments address the shortcomings of traditional compressor control systems in energy efficiency management, load regulation, and equipment stability through an IoT-based compressor control and management system. By real-time acquisition and analysis of multi-dimensional data, combined with long short-term memory networks and particle swarm optimization algorithms, this system can dynamically adjust the opening of the compressor's intake and exhaust valves, achieving precise control, optimizing compressor operating efficiency, extending equipment lifespan, and effectively reducing energy consumption.

[0018] The overall concept of the solution in this application embodiment is as follows: By connecting multi-dimensional sensors through the Internet of Things, multi-dimensional data and historical multi-dimensional data of the compressor can be obtained.

[0019] Based on multi-dimensional data and historical multi-dimensional data of the compressor, a mathematical relationship between the suction valve opening degree and the compressor load, energy efficiency and suction flow rate is established through a long short-term memory network. A suction valve model is constructed to predict the impact of suction valve opening degree adjustment on compressor efficiency under different loads.

[0020] Based on multi-dimensional data and historical multi-dimensional data of the compressor, a mathematical relationship between the opening degree of the exhaust valve and the compressor exhaust pressure, exhaust flow rate and compressor energy efficiency is established through support vector regression. An exhaust valve model is constructed to predict the impact of exhaust valve opening adjustment on compressor energy efficiency and stability under different load and exhaust pressure conditions.

[0021] Based on the intake valve model, combined with real-time load data and compressor energy efficiency requirements, the opening degree of the intake valve is dynamically adjusted through particle swarm optimization algorithm.

[0022] Based on the prediction model generated by the exhaust valve model building unit, and combined with real-time exhaust pressure, exhaust flow rate and compressor load data, the exhaust valve opening is dynamically adjusted through a fuzzy control algorithm.

[0023] Please see Figure 1This invention provides a technical solution: an IoT-based compressor control and management system, comprising the following modules: a multi-dimensional data acquisition module, a model building module, and a control module. The module is used to connect to multi-dimensional sensors via the IoT to acquire multi-dimensional data and historical multi-dimensional data of the compressor. The model building module includes an intake valve model building unit and an exhaust valve model building unit. The intake valve model building unit is used to establish a mathematical relationship between the intake valve opening degree and the compressor load, energy efficiency, and intake flow rate through a long short-term memory network based on the compressor's multi-dimensional data and historical multi-dimensional data, constructing an intake valve model to predict the impact of intake valve opening degree adjustment on compressor efficiency under different loads. The exhaust valve model building unit is used to establish a mathematical relationship between the intake valve opening degree and compressor load, energy efficiency, and intake flow rate based on the compressor's multi-dimensional data and historical multi-dimensional data, constructing an intake valve model to predict the impact of intake valve opening degree adjustment on compressor efficiency under different loads. The system utilizes dimensional data and historical multi-dimensional data to establish a mathematical relationship between the exhaust valve opening and the compressor's exhaust pressure, exhaust flow rate, and compressor energy efficiency through support vector regression. This constructs an exhaust valve model to predict the impact of exhaust valve opening adjustments on compressor energy efficiency and stability under different load and exhaust pressure conditions. The control module includes an intake valve control unit and an exhaust valve control unit. The intake valve control unit dynamically adjusts the intake valve opening based on the intake valve model, combined with real-time load data and compressor energy efficiency requirements, using a particle swarm optimization algorithm. The exhaust valve control unit dynamically adjusts the exhaust valve opening based on a prediction model generated by the exhaust valve model, combined with real-time exhaust pressure, exhaust flow rate, and compressor load data, using a fuzzy control algorithm.

[0024] In this implementation scheme, a multi-dimensional data acquisition module connects multiple multi-dimensional sensors via IoT technology to acquire real-time data related to compressor operation, including: compressor load (representing the compressor's current workload or load level); energy efficiency (the compressor's operating efficiency); suction flow rate (the compressor's intake flow rate, directly affecting its performance); discharge pressure (the pressure data on the compressor's discharge side); and discharge flow rate (the compressor's discharge flow rate). This data includes not only real-time acquisition but also historical data, used for model construction and training. The model construction module comprises two key parts: a suction valve model construction unit and a discharge valve model construction unit. The suction valve model construction unit uses Long Short-Term Memory (LSTM) networks to establish a mathematical relationship between the suction valve opening and compressor load, energy efficiency, and suction flow rate based on the acquired multi-dimensional and historical compressor data. LSTM is a neural network capable of capturing long-term dependencies in time-series data, effectively handling the impact of suction valve opening adjustments on compressor efficiency under different load and flow conditions. This model can accurately predict the optimization effect of suction valve opening on compressor performance. The exhaust valve model building unit uses the Support Vector Regression (SVR) algorithm to establish the mathematical relationship between the exhaust valve opening and the compressor's exhaust pressure, exhaust flow rate, and energy efficiency. The SVR algorithm excels at handling complex nonlinear regression problems and can analyze the impact of changes in the compressor's exhaust valve opening on the compressor's energy efficiency and stability based on historical data. By building the exhaust valve model, the exhaust valve opening can be predicted and optimized under different load and exhaust pressure conditions. The control module is the execution part of the entire system, responsible for adjusting the compressor's operating state in real time based on the model's results to improve its efficiency and stability. The control module includes two sub-modules: the intake valve control unit: This unit dynamically adjusts the intake valve opening using the Particle Swarm Optimization (PSO) algorithm, based on the intake valve model and combined with real-time acquired load data and compressor energy efficiency requirements. Particle Swarm Optimization is a swarm intelligence-based optimization algorithm that can quickly find the global optimum in a multi-dimensional space. By adjusting the intake valve opening, it enables the compressor to maintain optimal energy efficiency and performance under different loads. Exhaust Valve Control Unit: This unit uses a fuzzy control algorithm to dynamically adjust the opening of the exhaust valve based on the prediction results generated by the exhaust valve model, combined with real-time collected exhaust pressure, exhaust flow, and compressor load data. The fuzzy control algorithm has strong adaptability and robustness, enabling effective regulation under uncertain environments and ensuring stable compressor operation under complex load and exhaust pressure conditions. System Overall Workflow: Data Acquisition: Real-time and historical data of the compressor are collected via IoT devices. Model Training and Construction: Control models for the intake and exhaust valves are constructed using Long Short-Term Memory (LSTM) networks and Support Vector Regression (SVR) algorithms.Control Decision: The opening of the intake and exhaust valves is dynamically adjusted through particle swarm optimization and fuzzy control algorithms, thereby optimizing the performance and energy efficiency of the compressor.

[0025] Specifically, multi-dimensional data includes, but is not limited to, compressor load, exhaust temperature, exhaust pressure, intake volume, intake and exhaust valve opening, and motor speed.

[0026] In this implementation scheme, the compressor load refers to the degree of load borne by the compressor during operation, that is, the ratio between the actual compression capacity provided by the compressor and its design maximum capacity. The compressor load directly affects its energy efficiency and operational stability; therefore, real-time monitoring of load changes helps optimize control strategies and prevent equipment from operating under overload or inefficiently. Discharge temperature refers to the temperature of the gas at the compressor's discharge end. This data helps determine the thermodynamic state of the compressor during operation. Excessively high discharge temperatures may indicate low compressor efficiency, possibly due to factors such as excessively large intake valve opening or excessively small exhaust valve opening. Therefore, monitoring discharge temperature reflects the stability and energy efficiency of the compressor's performance. Discharge pressure refers to the pressure generated at the compressor's discharge end. Discharge pressure has a significant impact on compressor energy efficiency; excessively high or low discharge pressure may lead to decreased compressor efficiency or excessive wear. Therefore, by monitoring discharge pressure and using it in conjunction with other parameters, the compressor's operating state can be effectively controlled to maintain its optimal performance. Intake volume refers to the gas flow rate at the compressor's intake end. It directly affects the compressor's operating efficiency; excessively high or low intake volume will affect the compressor's load and energy efficiency. Under varying load conditions, a reasonable intake volume helps maintain the compressor's efficient operation; therefore, monitoring the intake volume is one of the important parameters for controlling compressor performance. Intake and Exhaust Valve Openings: The opening of the intake and exhaust valves affects the compressor's intake and exhaust flow rates, respectively. The intake valve opening determines the gas flow rate entering the compressor, while the exhaust valve opening affects the gas flow rate exiting. Dynamic adjustment of these valves directly impacts the compressor's compression efficiency and energy efficiency; reasonable valve opening adjustment helps optimize the compressor's operating state. Motor Speed: Motor speed refers to the speed at which the motor driving the compressor rotates, and it is usually related to the compressor's workload and operating efficiency. Excessively high or low motor speeds can lead to low compressor efficiency and may even affect equipment stability. Real-time monitoring of motor speed ensures that the motor operates within a suitable speed range, thereby maintaining system energy efficiency.

[0027] Specifically, the process of establishing the mathematical relationship between the intake valve opening and compressor load, energy efficiency, and intake flow rate through a Long Short-Term Memory (LSTM) network is as follows: Preprocessing is performed based on multi-dimensional compressor data and historical multi-dimensional data, including data cleaning, data normalization, and feature extraction, converting the data into a time series format. Historical data is divided into multiple samples according to a set time window, with each window serving as input data for the LTM network. Input data at each time step is segmented through windows, containing feature data on compressor load, energy efficiency, and intake flow rate. The network structure of the LTM model is set, including setting the input layer to accept feature data such as compressor load, energy efficiency, and intake flow rate; setting the number of network layers and the number of neurons in each layer to capture temporal dependencies between data; and setting the output layer to predict the intake valve opening. The LTM network model is trained using historical multi-dimensional data, including setting a loss function to measure the difference between predicted and actual values ​​and updating the model weights through backpropagation and gradient descent algorithms.

[0028] In this implementation plan, data preprocessing includes: Data cleaning: The cleaning process organizes the acquired multi-dimensional data, removing noise, missing values, and outliers to ensure data quality. Cleaned data is more accurate and suitable for subsequent modeling. Data normalization: Due to the significant differences in the dimensions and ranges of different data, normalization compresses the data to a uniform scale, typically through standardization or min-max normalization. This avoids excessively large ranges for certain feature values, which could affect the training efficiency and accuracy of the model. Feature extraction: Feature extraction extracts meaningful features that describe system behavior from the raw data. For example, features related to compressor load, energy efficiency, and intake flow rate are extracted from the raw data for subsequent modeling and analysis. Time series format: Converting historical multi-dimensional data into a time series format captures the temporal relationships of the data, enabling the model to predict future states based on historical input data. In time series data, each data point represents the state at a specific timestamp. Historical data is divided into multiple samples according to a set time window. Each time window serves as the input data for the LSTM network. For example, if the time window is set to 10 seconds, data from every 10 seconds will be input into the LSTM model as a sample. The data within each time window contains multiple features, such as compressor load, energy efficiency, and intake air flow. These features are used as input at each time step of the model, helping the model learn the temporal changes of each feature. Input Layer: The input layer receives data from each time window, including features related to compressor load, energy efficiency, and intake air flow. Each input feature is mapped to a neuron in the input layer, providing the input signal for subsequent network computations. Number of Layers and Neurons: The number of layers and neurons in each layer of the LSTM model determines the model's complexity and expressive power. Multi-layered networks can better capture the complex relationships in the data, while the number of neurons in each layer affects the model's learning ability. In each layer, the LSTM uses a gating mechanism to capture the temporal dependencies of the data. Output Layer: The output layer predicts the opening of the intake valve. After network training, the output layer provides a numerical value representing the predicted opening of the intake valve under the current input conditions. Loss Function: The loss function measures the difference between the model's predictions and the true results. Common loss functions include mean squared error (MSE). Optimizing the loss function is the core of the entire training process, aiming to improve the model's accuracy by minimizing the prediction error. Backpropagation and Gradient Descent: Backpropagation calculates the gradient of the loss function with respect to model parameters (such as weights and biases), and then gradient descent updates these parameters, gradually optimizing the model. Backpropagation helps the model adjust weights, making the predicted values ​​closer to the true values, thereby improving model performance.

[0029] Specifically, the process of constructing the intake valve model is as follows: the mathematical relationship between the intake valve opening and the compressor load, energy efficiency and intake flow is obtained through a trained long short-term memory network model; based on the obtained mathematical relationship, combined with the real-time collected compressor load, energy efficiency and intake flow, the intake valve opening under different conditions is calculated, and an intake valve model between intake valve opening adjustment and compressor efficiency is generated.

[0030] In this implementation plan, the process of constructing the intake valve model involves: obtaining the mathematical relationship between the intake valve opening and compressor load, energy efficiency, and intake flow rate. Through a trained Long Short-Term Memory (LSTM) network model, the complex temporal relationship between the intake valve opening and compressor load, energy efficiency, and intake flow rate can be learned from historical data. The learning relationship is defined as follows: in: Indicates time The opening degree of the intake valve. Load Indicates time The compressor load under the specified conditions. Efficiency Indicates time The compressor's energy efficiency is as follows. (SuctionFlow) Indicates time The lower inhalation flow rate. This indicates that the nonlinear mapping relationship obtained through the trained LSTM model captures the dependency between compressor load, energy efficiency, and suction flow rate. The suction valve opening under different conditions can be dynamically calculated based on the relationship obtained from the LSTM model and combined with real-time collected compressor load, energy efficiency, and suction flow rate data. The formula for calculating the suction valve opening is as follows: Where W and b represent the weights and biases of the LSTM model, respectively, which are parameters obtained through model training. g() represents the prediction function of the trained LSTM model, used to calculate the intake valve opening based on real-time data. After obtaining the intake valve opening, the relationship between intake valve opening adjustment and compressor efficiency can be further analyzed. This can be expressed as: Among them: Efficiency Indicates time The compressor's energy efficiency is as follows. The intake valve opening is calculated using the formula above, where c is a parameter of the model obtained by fitting historical data. This is a function representing the relationship between the suction valve opening and the compressor efficiency. Combining these relationships yields a complete suction valve model, used to calculate the suction valve opening based on real-time load, energy efficiency, and suction flow rate, thereby predicting and optimizing compressor efficiency. The final suction valve model can be expressed as: ; This model enables the dynamic adjustment of the suction valve opening based on real-time collected compressor data (load, energy efficiency, suction flow), thereby optimizing the compressor's operating efficiency and stability.

[0031] Specifically, the process of establishing the mathematical relationship between the exhaust valve opening and the compressor exhaust pressure, exhaust flow rate, and compressor energy efficiency through support vector regression is as follows: Based on historical multi-dimensional data, the historical data is preprocessed using a support vector regression model, including data cleaning, normalization, and feature extraction; the training data is fitted using the support vector regression model to establish the mathematical relationship between the exhaust valve opening and the compressor exhaust pressure, exhaust flow rate, and energy efficiency; and the accuracy and generalization ability of the model are ensured by optimizing the hyperparameters in the SVR model, including the penalty factor and kernel function type.

[0032] In this implementation plan, data preprocessing involves the following steps: First, historical multi-dimensional data is preprocessed to ensure data quality and adaptability. This process includes: Data cleaning: removing noisy data and missing values ​​to ensure the accuracy and completeness of the data used for modeling. Normalization: to improve model training performance, the data is scaled proportionally to ensure all feature data are on the same order of magnitude, avoiding excessive influence of certain features on model training. Feature extraction: extracting useful features for prediction from the raw data. These features may include exhaust pressure, exhaust flow rate, and compressor energy efficiency, providing valuable input to the model. Model fitting and training: using a Support Vector Regression (SVR) model, the model is fitted and trained using historical data. SVR is a regression analysis method based on statistical learning theory that effectively captures the nonlinear relationship between input features and the target variable (exhaust valve opening). During training, SVR optimizes the objective function to find the most suitable regression function for the data, thereby obtaining the mathematical relationship between exhaust valve opening and compressor exhaust pressure, exhaust flow rate, and energy efficiency. Hyperparameter optimization: To improve the accuracy and generalization ability of the SVR model, its hyperparameters need to be optimized. This mainly includes: Penalty factor (C): This parameter controls the model's error tolerance during the fitting process. A larger C value reduces training error but may lead to overfitting; a smaller C value increases error tolerance and helps avoid overfitting. Kernel function type: SVR models use different kernel functions (such as linear kernels, radial basis kernels, etc.) to handle the nonlinear relationships in the input data. Choosing an appropriate kernel function can improve the model's predictive ability and ensure that the model better captures the complex relationship between the exhaust valve opening and other variables. Model evaluation and validation: The trained SVR model is evaluated through cross-validation or independent validation sets to ensure its stable performance on different datasets and avoid overfitting or underfitting.

[0033] Specifically, the process of constructing the exhaust valve model is as follows: A mathematical relationship between the exhaust valve opening and the compressor's exhaust pressure, exhaust flow rate, and energy efficiency is obtained through a trained support vector regression model; based on the obtained mathematical relationship, and combined with real-time collected data on the compressor's exhaust pressure, exhaust flow rate, and energy efficiency, the exhaust valve opening under different operating conditions is calculated; an exhaust valve model is generated that relates the exhaust valve opening adjustment to the compressor's energy efficiency and stability, used to dynamically adjust the compressor's exhaust valve opening and optimize the compressor's operating efficiency and stability.

[0034] In this implementation plan, during the specific process of constructing the exhaust valve model, a trained Support Vector Regression (SVR) model is used to obtain the mathematical relationship between the exhaust valve opening and the compressor exhaust pressure, exhaust flow rate, and energy efficiency. The detailed process is as follows: Model Training and Mathematical Relationship Acquisition: Historical data is input into the training model using the Support Vector Regression model to obtain the exhaust valve opening. With compressor discharge pressure Exhaust flow rate and energy efficiency The mathematical relationship between them. After training, the SVR model can derive a mapping function; This function describes the variation of the exhaust valve opening under different discharge pressure, discharge flow rate, and energy efficiency conditions of the compressor. The exhaust valve opening is calculated using real-time data: Based on a trained SVR model, during actual operation, the exhaust valve opening under different operating conditions is calculated using the aforementioned mathematical relationships, combined with real-time collected compressor discharge pressure, discharge flow rate, and energy efficiency data. The specific calculation process is as follows: ;in, This indicates the current time point, and the opening degree of the exhaust valve is continuously updated and calculated using real-time data. An exhaust valve model is generated to optimize compressor operating efficiency and stability: By obtaining the relationship between the exhaust valve opening degree and compressor energy efficiency and stability, an exhaust valve model is further constructed to dynamically adjust the compressor's exhaust valve opening degree, thereby optimizing the compressor's operating efficiency and stability. This model not only considers energy efficiency (… The improvement also considers stability factors, such as the stability of exhaust pressure. By adjusting the exhaust valve opening in real time, the compressor's operating state can be effectively controlled, avoiding over-compression or inefficient operation, ultimately achieving energy savings and improving system stability. The key objective of the exhaust valve model is: ;in, The function, fitted by a support vector regression (SVR) model, describes the relationship between the discharge valve opening and the compressor's state (discharge pressure, discharge flow rate, and energy efficiency). The goal is to ensure the compressor operates in a highly efficient and stable state by dynamically adjusting the discharge valve opening.

[0035] Specifically, the process of dynamically adjusting the intake valve opening using the particle swarm optimization algorithm is as follows: Real-time compressor load, energy efficiency, and intake flow data are acquired and input into the intake valve model to calculate the current intake valve opening. Based on the real-time data and compressor energy efficiency requirements, a target intake valve opening is set, and the difference between the target opening and the current opening is calculated. The particle swarm in the particle swarm optimization algorithm is initialized, and a fitness function is set for each particle. This function measures how close the adjusted intake valve opening is to the target opening. The particle positions are iteratively updated using the particle swarm optimization algorithm, and the fitness function is calculated in each iteration. The particle with the best fitness is selected as the optimal solution. When the iterative process of the particle swarm converges, the optimal solution is the optimized intake valve opening, used to dynamically adjust the intake valve and optimize the compressor's operating efficiency.

[0036] In this implementation scheme, real-time data acquisition and suction valve model calculation of the current valve opening are performed as follows: First, the system collects real-time data on compressor load, energy efficiency, and suction flow. This real-time data is input to the suction valve model, which calculates the current suction valve opening based on this data. The purpose of this step is to determine the current suction valve opening based on the real-time monitored operating status and to serve as the starting point for the optimization process. A target opening is set and the difference is calculated: Based on the compressor's actual load and energy efficiency requirements, a target suction valve opening is set. The target opening is set based on the system's optimal operating efficiency and energy efficiency requirements. The difference between the current opening and the target opening is calculated. This difference will be used in the fitness function of the particle swarm optimization algorithm to guide the direction of algorithm optimization. The target opening can be dynamically adjusted based on real-time data such as compressor load, energy efficiency, and suction flow to ensure that the compressor always operates in an optimal state. The particle swarm optimization algorithm is initialized: In the particle swarm optimization algorithm, a group of particles is first initialized. Each particle represents a possible suction valve opening solution. Each particle has a position and a velocity. Position represents the particle's current solution (i.e., the current intake valve opening), and velocity determines how quickly the particle updates its position in the search space. The size of the particle swarm depends on the complexity of the problem; generally, a larger number of particles allows for a wider search range, but also increases computational complexity. Define a fitness function and calculate the particle's fitness: For each particle, define a fitness function that measures how close the particle's current solution is to the target solution. The fitness function is typically the error between the target valve opening and the particle's position (current intake valve opening); a smaller error results in a higher fitness. The fitness function can be designed as follows: ;in, It is the target opening. This is the current aperture of the particle. The fitness function aims to minimize this difference to obtain the optimal solution. Particle Swarm Optimization Iterative Update: The Particle Swarm Optimization algorithm iteratively updates the velocity and position of particles, gradually approaching the global optimum. In each iteration, the particle's velocity and position are adjusted based on its own experience (individual optimal position) and the swarm's experience (swarm optimal position). ; ;in, For particles At time step speed, For particles Position (i.e., current intake valve opening) For particles The best historical location This is the best position for the entire group. As a learning factor, It is a random number. The inertia weights are used. Through continuous iteration, the particles gradually converge to the optimal solution, which is the value closest to the target intake valve opening. In each iteration, the particle swarm gradually reduces the opening difference through optimization until the algorithm converges. Convergence usually means that the value of the fitness function no longer changes significantly, or that the set maximum number of iterations has been reached. When the particle swarm converges, the solution corresponding to the best particle is the optimized intake valve opening, which makes the intake valve adjustment best meet the compressor's energy efficiency requirements. Finally, the optimized intake valve opening is dynamically adjusted in the compressor, thereby optimizing the compressor's operating efficiency.

[0037] Specifically, the process of dynamically adjusting the exhaust valve opening using a fuzzy control algorithm is as follows: Real-time data on exhaust pressure, exhaust flow rate, and compressor load are acquired and used as input variables for the fuzzy control algorithm. Based on the actual data of exhaust pressure, exhaust flow rate, and compressor load, the input variables are transformed into fuzzy sets through fuzzification processing. A fuzzy control rule base is established, and based on the exhaust valve model, the relationship between exhaust pressure, exhaust flow rate, and compressor load is determined to determine the direction and extent of exhaust valve opening adjustment. Fuzzy inference is used to reason about the fuzzy control rules, calculating the adaptive output value of the exhaust valve, i.e., the exhaust valve opening adjustment amount. The fuzzy inference result is defuzzified to obtain the actual exhaust valve opening adjustment value, and the exhaust valve opening is dynamically adjusted through the control system to optimize the compressor's energy efficiency and stability.

[0038] In this implementation plan, real-time data acquisition is performed as follows: First, the system collects data related to the compressor's operating status in real time, mainly including discharge pressure, discharge flow rate, and compressor load. This data serves as the input variables for the fuzzy control algorithm, determining the adjustment of the discharge valve opening. Fuzzification of input variables: Fuzzification is the process of converting precise input data into fuzzy sets. In this step, the three input variables—discharge pressure, discharge flow rate, and compressor load—are converted into fuzzy values ​​according to a preset fuzzy membership function. Each input variable is typically divided into several different linguistic variables ("high," "medium," and "low") and mapped to fuzzy sets. For example, discharge pressure can be divided into three fuzzy states: "high pressure," "medium pressure," and "low pressure," and compressor load can be divided into "heavy load," "medium load," and "light load." This fuzzification process helps to address noise and uncertainty in the input data, enabling the fuzzy control system to flexibly handle complex, nonlinear control problems. Setting up a fuzzy control rule base: The fuzzy rule base is one of the core components of the fuzzy control system. It contains a set of rules describing the relationship between input and output variables. In this process, a set of control rules based on actual operating requirements needs to be set according to the discharge valve model. Control rules typically consist of "if...then..." statements, such as: "If the discharge pressure is high and the compressor load is heavy, then the discharge valve opening should be significantly reduced." These rules reflect the relationship between discharge pressure, discharge flow rate, and compressor load under different operating conditions, and indicate the direction and extent of discharge valve opening adjustment. Fuzzy Inference: Fuzzy inference is the process of deriving fuzzy control rules. In this step, the control system calculates the adjustment amount of the discharge valve opening through inference to determine the output adjustment amount corresponding to each input state. The inference process considers the influence of all fuzzy rules and combines them to obtain a discharge valve adjustment amount that conforms to the actual situation. Defuzzification: After fuzzy inference, the output value obtained by the system is still fuzzy and needs to be converted into a precise discharge valve opening adjustment value through the defuzzification process. Defuzzification methods include the central averaging method and the weighted averaging method, which can map the fuzzy output to the actual numerical range. The result obtained after defuzzification is the actual discharge valve opening adjustment amount, which will be used to actually adjust the discharge valve opening in the control system. Dynamic adjustment of exhaust valve opening: Finally, the obtained exhaust valve opening adjustment value will be dynamically adjusted by the control system. The exhaust valve opening will be adjusted based on real-time data (such as exhaust pressure, exhaust flow, and compressor load) and the adjustment amount calculated by the fuzzy control algorithm. Through continuous dynamic adjustment, the compressor's exhaust valve opening is ensured to always be in the optimal state under different operating conditions, so as to maximize the compressor's energy efficiency and maintain its stability.Optimizing compressor energy efficiency and stability: Through continuous adjustment using a fuzzy control algorithm, the discharge valve opening can respond in real time to factors such as compressor load changes and discharge pressure fluctuations, thereby optimizing the compressor's operating status. Specifically, by properly adjusting the discharge valve, compressor energy efficiency can be improved, energy consumption reduced, and operational stability enhanced, while avoiding operational instability or low energy efficiency caused by excessively high or low opening degrees.

[0039] In summary, this application has at least the following effects: The IoT-based compressor control and management system dynamically adjusts the opening of the intake and exhaust valves to precisely control the compressor's operating status, improving energy efficiency, reducing energy consumption, and maximizing energy utilization. Precise valve opening adjustment balances changes in compressor load and exhaust pressure, preventing equipment vibration or malfunctions caused by improper valve control, thus enhancing the compressor's long-term stability and safety. Employing Long Short-Term Memory (LSTM) networks, Support Vector Regression (SVR), Particle Swarm Optimization (PSO), and fuzzy control algorithms, the system achieves intelligent dynamic adjustment of the compressor opening, reducing human intervention and operational complexity. It can flexibly respond to various changes and disturbances in complex and uncertain environments, improving the system's anti-interference capabilities and ensuring efficient compressor operation. Through automated and intelligent control methods, reliance on manual operation is reduced, operating procedures are simplified, the technical threshold for operators is lowered, and the convenience and accuracy of operation are improved.

[0040] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0041] This invention is described with reference to flowchart illustrations and / or block diagrams of systems, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0042] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0043] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0044] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0045] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. The compressor control management system based on the Internet of Things is characterized by: It includes the following modules: multi-dimensional data acquisition module, model building module, and control module; The module is used to connect the multi-dimensional sensor through the Internet of Things to obtain the multi-dimensional data and historical multi-dimensional data of the compressor; The model building module includes an intake valve model building unit and an exhaust valve model building unit; The suction valve model building unit is used to establish a mathematical relationship between the suction valve opening and the compressor load, energy efficiency and suction flow rate through a long short-term memory network according to the multi-dimensional data and historical multi-dimensional data of the compressor, so as to build a suction valve model; The exhaust valve model building unit is used to establish a mathematical relationship between the exhaust valve opening and the exhaust pressure, exhaust flow rate and energy efficiency of the compressor through support vector regression according to the multi-dimensional data and historical multi-dimensional data of the compressor, so as to build an exhaust valve model; The control module is used to include an intake valve control unit and an exhaust valve control unit; The suction valve control unit is used to dynamically adjust the opening of the suction valve through a particle swarm optimization algorithm according to the suction valve model, combined with real-time load data and compressor energy efficiency requirements; The exhaust valve control unit is used to dynamically adjust the exhaust valve opening through a fuzzy control algorithm based on the prediction model generated by the exhaust valve model building unit in combination with real-time exhaust pressure, exhaust flow and compressor load data.

2. The compressor control management system based on the Internet of Things according to claim 1, characterized in that: The multi-dimensional data includes but is not limited to compressor load, exhaust temperature, exhaust pressure, intake volume, intake valve and exhaust valve opening, and motor speed.

3. The compressor control management system based on the Internet of Things according to claim 2, characterized in that: The specific process of establishing the mathematical relationship between the suction valve opening and the compressor load, energy efficiency and suction flow rate through the long short-term memory network is as follows: Based on the multi-dimensional data of the compressor and the historical multi-dimensional data, pre-processing is performed, including: data cleaning, data normalization and feature extraction, and the data is converted into a time series format; The historical data is divided into multiple samples according to the set time window, and each window is used as the input data of the long short-term memory network; The input data of each time step is divided by windows, and the input data includes characteristic data of compressor load, energy efficiency and suction flow rate; Setting the network structure of the long short-term memory network model, including setting the input layer to receive characteristic data such as compressor load, energy efficiency and suction flow, setting the number of network model layers and the number of neurons in each layer to capture the temporal dependency between data, and setting the output layer to predict the opening of the suction valve; The long short-term memory network model is trained through historical multi-dimensional data, including setting the loss function to measure the difference between the predicted value and the actual value and updating the model weights through back propagation and gradient descent algorithms.

4. The compressor control management system based on the Internet of Things according to claim 3 is characterized in that: The specific process of building the suction valve model is as follows: The mathematical relationship between the suction valve opening and the compressor load, energy efficiency and suction flow rate is obtained through the trained long short-term memory network model; Based on the acquired mathematical relationship, combined with the real-time collected compressor load, energy efficiency, and intake flow rate, the intake valve opening under different conditions is calculated, and an intake valve model between the intake valve opening adjustment and the compressor efficiency is generated. This is used to predict the impact of the intake valve opening adjustment on the compressor efficiency under different loads.

5. The compressor control management system based on the Internet of Things according to claim 4 is characterized in that: The specific process of establishing the mathematical relationship between the exhaust valve opening and the compressor exhaust pressure, exhaust flow rate and compressor energy efficiency through support vector regression is as follows: Through the support vector regression model, historical data is preprocessed based on historical multi-dimensional data, including data cleaning, normalization and feature extraction; The mathematical relationship between the exhaust valve opening and the compressor exhaust pressure, exhaust flow rate and energy efficiency is established by fitting the training data with the support vector regression model; The accuracy and generalization ability of the model are ensured by optimizing the hyperparameters in the SVR model, including the penalty factor and kernel function type.

6. The compressor control management system based on the Internet of Things according to claim 5, characterized in that: The specific process of building the exhaust valve model is as follows: The mathematical relationship between the exhaust valve opening and the compressor exhaust pressure, exhaust flow rate and energy efficiency is obtained through the trained support vector regression model; Based on the obtained mathematical relationship, combined with the real-time collected data such as compressor exhaust pressure, exhaust flow rate and energy efficiency, the exhaust valve opening under different working conditions is calculated; An exhaust valve model between exhaust valve opening adjustment and compressor energy efficiency and stability is generated to predict the impact of exhaust valve opening adjustment on compressor energy efficiency and stability under different load and exhaust pressure conditions.

7. The compressor control management system based on the Internet of Things according to claim 6, characterized in that: The specific process of dynamically adjusting the opening of the intake valve through the particle swarm optimization algorithm is as follows: Obtain the real-time collected compressor load, energy efficiency and suction flow data, and input them into the suction valve model to calculate the current suction valve opening; According to the real-time collected data and the compressor energy efficiency requirements, the target opening of the suction valve is set, and the difference between the target opening and the current opening is calculated; Initialize the particle swarm in the particle swarm optimization algorithm and set the fitness function of each particle, which measures the closeness between the opening of the intake valve adjusted by the particle and the target opening; The particle positions are iteratively updated through the particle swarm optimization algorithm, and the fitness function is calculated in each iteration, and the particle with the best fitness is selected as the best solution; When the iterative process of the particle swarm converges, the best solution obtained is the optimized suction valve opening, which is used to dynamically adjust the suction valve and optimize the operating efficiency of the compressor.

8. The compressor control management system based on the Internet of Things according to claim 7, characterized in that: The specific process of dynamically adjusting the exhaust valve opening through the fuzzy control algorithm is as follows: Obtain real-time collected exhaust pressure, exhaust flow and compressor load data as input variables of the fuzzy control algorithm; According to the actual data of exhaust pressure, exhaust flow and compressor load, the input variables are transformed into fuzzy sets through fuzzy processing; Set up the fuzzy control rule base, determine the relationship between exhaust pressure, exhaust flow and compressor load according to the exhaust valve model, and determine the direction and degree of adjustment of the exhaust valve opening; The fuzzy control rules are inferred through fuzzy reasoning to calculate the adaptive output value of the exhaust valve, that is, the opening adjustment amount of the exhaust valve; The fuzzy reasoning result is defuzzified to obtain the actual exhaust valve opening adjustment value, and the exhaust valve opening is dynamically adjusted through the control system to optimize the energy efficiency and stability of the compressor.

Images