Intelligent control system of air compressor

By introducing machine learning technology and multi-module collaborative work in the air compressor control system, the problems of traditional air compressor control systems in load distribution, dynamic characteristics understanding and human-computer interaction are solved, and efficient, stable and safe intelligence is achieved Control improves energy utilization efficiency and system stability.

CN119024745APending Publication Date: 2024-11-26GUANGDONG HANDE PRECISION MASCH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411123108.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-15
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

Traditional air compressor control systems are difficult to achieve optimal load distribution when multiple air compressors work together, have low energy utilization efficiency, lack in-depth understanding of dynamic characteristics, and are difficult to adapt to complex and changeable gas needs, affecting system stability. Moreover, the human-computer interaction interface is poor, making it difficult for operators to intuitively and quickly grasp the system status and make decisions.

Method used

An intelligent control system for air compressors was designed, including data acquisition module, central controller, machine learning module, collaborative control module, execution module, human-computer interaction module and safety monitoring module, using pressure sensors, flowmeters, and power analyzers , temperature sensors and vibration sensors collect data, build dynamic operation models and learn optimal control strategies through machine learning modules, collaborative control modules determine the operating parameters of each air compressor, perform modules to adjust the working status, and human-computer interaction modules display system status And receive user input, the security monitoring module monitors exceptions and triggers the protection mechanism.

Benefits of technology

It significantly improves the energy utilization efficiency and operation stability of the system, realizes self-optimization and forward-looking control, improves the operability and reliability of the system, and uses an intuitive and friendly interactive interface and a complete security monitoring mechanism.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119024745A_ABST
    Figure CN119024745A_ABST
Patent Text Reader

Abstract

The invention relates to the field of industrial automation control, in particular to an intelligent control system for air compressors, which comprises a data acquisition module used for acquiring operation parameters of a plurality of air compressors; the central controller is electrically connected with the data acquisition module and is used for receiving the operation parameters and generating a control instruction based on the operation parameters; the machine learning module is arranged in the central controller and used for constructing an air compressor dynamic operation model and learning an optimal control strategy based on the operation parameters; the cooperative control module is electrically connected with the central controller and used for determining operation parameters of each air compressor based on the optimal control strategy; the execution module is electrically connected with the cooperative control module and used for adjusting the working state of each air compressor based on the operation parameters; the man-machine interaction module is electrically connected with the central controller and is used for displaying the running state of the system and receiving user input; and the safety monitoring module is electrically connected with the central controller and is used for monitoring the operation safety of the system and triggering a protection mechanism when abnormity is detected.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of industrial automation control, and in particular to an intelligent control system for an air compressor based on machine learning. Background Art

[0002] As an important power equipment in industrial production, the efficient and stable operation of air compressors is crucial to the production process. With the advent of the Industry 4.0 era, traditional air compressor control systems have gradually revealed many shortcomings. First, when multiple air compressors work together, existing systems often have difficulty achieving optimal load distribution, resulting in low energy efficiency. Second, due to a lack of in-depth understanding of the dynamic characteristics of the air compressor system, existing control strategies are difficult to adapt to complex and changing gas demand, affecting system stability. Furthermore, the existing system's insufficient mining and utilization of historical operating data makes it difficult to achieve continuous optimization and self-learning. Finally, the human-computer interaction interface is poorly designed, making it difficult for operators to intuitively and quickly grasp the system status and make reasonable decisions. Summary of the Invention

[0003] In view of the above problems, the present invention aims to provide an intelligent air compressor control system to achieve intelligent coordinated control of multiple air compressors, significantly improving the system energy efficiency and operational stability.

[0004] To this end, the present invention adopts the following technical solution: an air compressor intelligent control system, comprising:

[0005] Data acquisition module, used to obtain the operating parameters of multiple air compressors;

[0006] a central controller, electrically connected to the data acquisition module, configured to receive the operating parameters and generate control instructions based on the operating parameters;

[0007] a machine learning module, provided in the central controller, for constructing a dynamic operation model of the air compressor and learning an optimal control strategy based on the operating parameters;

[0008] a collaborative control module, electrically connected to the central controller, for determining operating parameters of each air compressor based on the optimal control strategy;

[0009] an execution module, electrically connected to the collaborative control module, and configured to adjust the working state of each air compressor based on the operating parameters;

[0010] A human-computer interaction module, electrically connected to the central controller, for displaying the system operating status and receiving user input; and

[0011] The safety monitoring module is electrically connected to the central controller and is used to monitor the safety of system operation and trigger a protection mechanism when an anomaly is detected.

[0012] Specifically, the data acquisition module includes:

[0013] The pressure sensor, flow meter, power analyzer, temperature sensor and vibration sensor are used to obtain the pressure, flow, power, temperature and vibration data of the air compressor respectively.

[0014] Specifically, the central controller includes:

[0015] A processor, a memory and a communication interface, wherein the processor is used to execute instructions stored in the memory, and the communication interface is used to exchange data with other modules.

[0016] Specifically, the machine learning module includes:

[0017] A data preprocessing submodule, for cleaning, normalizing and feature extracting the operating parameters;

[0018] An air compressor model building submodule, configured to build a dynamic operation model of each air compressor based on the preprocessed operation parameters;

[0019] a reinforcement learning submodule, configured to learn an optimal control strategy based on the dynamic operation model; and

[0020] The demand forecast submodule is used to predict the gas demand in the future.

[0021] Specifically, the air compressor model construction submodule uses a recurrent neural network to construct the dynamic operation model.

[0022] Specifically, the reinforcement learning submodule adopts a deep Q-network algorithm to learn the optimal control strategy.

[0023] Specifically, the demand prediction submodule uses a long short-term memory network to predict the gas demand.

[0024] Specifically, the collaborative control module adopts a model predictive control framework to determine the operating parameters of each air compressor based on the optimal control strategy and the gas demand.

[0025] Specifically, it also includes:

[0026] Edge computing nodes are located near each air compressor to execute real-time control logic and data preprocessing.

[0027] Specifically, it also includes:

[0028] The heat recovery module is electrically connected to the central controller and is used to monitor the heat discharged by the air compressor and optimize the operation strategy of the air compressor based on the heat energy utilization efficiency.

[0029] The air compressor intelligent control system primarily consists of a data acquisition module, a central controller, a machine learning module, a collaborative control module, an execution module, a human-computer interaction module, and a safety monitoring module. The data acquisition module is responsible for collecting real-time data on air compressor operating parameters, such as pressure, flow, power, temperature, and vibration. This data is transmitted to the central controller via industrial Ethernet, where it is deeply analyzed and processed by the machine learning module.

[0030] The machine learning module is the core of the system, consisting of four submodules: data preprocessing, air compressor model building, reinforcement learning, and demand forecasting. Through the collaborative work of these submodules, the system can continuously learn and optimize control strategies to adapt to various complex operating conditions.

[0031] Based on the output of the machine learning module, the collaborative control module calculates and generates the optimal operating parameters for each air compressor. These parameters are then implemented by the execution module to adjust the operating status of each air compressor. To facilitate operator understanding of system operation, the present invention also features an intuitive and user-friendly human-computer interaction interface that displays system status, energy efficiency analysis results, and optimization suggestions in real time.

[0032] In addition, considering the safety needs of industrial sites, this system is also equipped with a safety monitoring module that can monitor the system operation status in real time and trigger the protection mechanism in time when an anomaly is detected to ensure the safety of equipment and personnel.

[0033] The advantages of this invention are primarily reflected in the following aspects: First, by introducing advanced machine learning technology, intelligent control of the air compressor system is achieved, significantly improving energy efficiency. Second, the use of a reinforcement learning algorithm enables the system to self-optimize and continuously adapt to complex and changing operating conditions. Third, combined with a demand forecasting model, proactive control of the air compressor is achieved, effectively improving system stability. Finally, a user-friendly interactive interface and comprehensive safety monitoring mechanisms greatly enhance the system's operability and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is the overall framework diagram of the air compressor intelligent control system of the present invention.

[0035] Figure 2 Shows the internal structure of the data acquisition module.

[0036] Figure 3 This is the functional block diagram of the machine learning module.

[0037] Figure 4 The workflow of the collaborative control module is described.

[0038] Figure 5 The interface diagram of the human-computer interaction module is given. DETAILED DESCRIPTION

[0039] The specific embodiments of the present invention will be described in more detail below with reference to the accompanying drawings.

[0040] like Figure 1 As shown, the air compressor intelligent control system of the present invention mainly consists of seven modules: data acquisition module 1, central controller 2, machine learning module 3, collaborative control module 4, execution module 5, human-computer interaction module 6, and safety monitoring module 7. These modules are closely connected via industrial Ethernet to form a complete closed-loop control system.

[0041] First, let's look at the specific implementation of data acquisition module 1. Figure 2 As shown, this module includes multiple sensors for comprehensively collecting operating data from the air compressor. Pressure sensor 11 utilizes a high-precision piezoresistive design with a measurement range of 0-1.6 MPa and an accuracy of ±0.1% FS. Flowmeter 12 utilizes a thermal mass flowmeter, which directly measures flow under standard operating conditions with an accuracy of ±1% FS. Power analyzer 13 features a wideband capability, accurately measuring power quality parameters within the 0.5 Hz-100 kHz range. Temperature sensor 14 utilizes a PT100 platinum resistance resistor with a measurement range of -50°C to 200°C and an accuracy of ±0.1°C. Vibration sensor 15 utilizes a triaxial MEMS accelerometer with a frequency response range of 0.5 Hz-10 kHz. These sensors are connected to the local data acquisition unit 16 via RS485 or 4-20 mA signals. Notably, this acquisition unit utilizes an industrial-grade ARM processor and supports edge computing, enabling local processing of some data preprocessing, effectively reducing the burden on the central controller.

[0042] Next, we will introduce the core of the system - machine learning module 3 in detail. Figure 3 As shown, this module includes four submodules: data preprocessing submodule 31, air compressor model building submodule 32, reinforcement learning submodule 33 and demand forecasting submodule 34. These submodules work closely together to achieve intelligent control of the system.

[0043] The data preprocessing submodule 31 first cleans and smoothes the collected raw data. To remove outliers, we use a median filter algorithm; for missing data, linear interpolation is used to complete the data. Subsequently, the data is smoothed using an exponentially weighted moving average (EWMA) method based on a sliding window to eliminate the impact of short-term fluctuations. On this basis, we also perform feature engineering to extract time features (such as hours, dates, and days of the week), statistical features (such as mean, standard deviation, skewness, kurtosis, etc.), and frequency domain features (mainly performing fast Fourier transform on vibration data). These carefully processed data lay a solid foundation for subsequent model training.

[0044] The air compressor model building submodule 32 uses a bidirectional gated recurrent unit network (Bi-GRU) to capture the dynamic characteristics of the air compressor. The reason for choosing Bi-GRU instead of traditional RNN or LSTM is that it can consider both past and future information simultaneously, making it more suitable for processing time series data. The network structure is as follows: the input layer receives normalized multi-dimensional time series data, followed by two Bi-GRU layers (128 hidden units in the first layer and 64 hidden units in the second layer), followed by a fully connected layer of 32 neurons, and finally the output layer predicts the operating status in the short term in the future. To prevent overfitting, we used Dropout technology and L2 regularization during the training process.

[0045] The reinforcement learning submodule 33 is the decision-making core of this system and adopts the Double DQN with Prioritized Experience Replay algorithm. Compared with traditional Q-learning, this algorithm can effectively solve the over-estimation problem and improve learning efficiency through prioritized experience replay. When defining the state space, we not only consider the operating parameters of each air compressor (such as pressure, flow, power, etc.), but also include the overall state of the system (such as total pressure, total flow, pipe network pressure loss, etc.) and environmental factors (such as ambient temperature, humidity, etc.). The action space includes the on / off status of the air compressor, the output frequency of the inverter and the pressure set point. The design of the reward function comprehensively considers factors such as energy saving, pressure stability, load balance, and the number of equipment starts and stops, and balances different goals by adjusting the weights of each item.

[0046] The demand forecasting submodule 34 uses the Temporal Fusion Transformers (TFT) model, an advanced model that combines LSTM, self-attention mechanism and time perception layer. Compared with traditional time series forecasting methods, TFT can better handle multivariate time series data and has strong interpretability. The input of the model includes historical gas consumption data (hourly gas consumption in the last 7 days), time features (hour, date, week, month, holiday sign), production plan (planned output in the next 24 hours) and weather forecast data (temperature, humidity, and air pressure forecast values). With these rich input features, the model can accurately predict the hourly gas consumption and its confidence interval for the next 24 hours.

[0047] Based on the output of the machine learning module, the collaborative control module 4 (such as Figure 4 (as shown) uses a model predictive control (MPC) framework to generate the optimal control strategy. The advantage of MPC is that it simultaneously considers current conditions and future forecasts, enabling proactive control. To set the optimization objectives, we employed a multi-objective optimization approach, comprehensively considering multiple goals, including minimizing energy consumption, minimizing pressure fluctuations, and balancing equipment wear. By adjusting the weights of each objective, we can flexibly adapt to different application scenarios and user needs.

[0048] In order to facilitate operators to grasp the system status in real time and make timely decisions, we have carefully designed the human-computer interaction module 6 (such as Figure 5 (As shown). This module utilizes a responsive web design and supports multi-terminal access. The interface not only displays the operating parameters of each air compressor and the overall system status in real time, but also provides a wealth of data visualization features, such as energy consumption trend charts and pressure fluctuation graphs. Furthermore, the system automatically generates optimization suggestions based on current status and historical data to assist operators in making decisions.

[0049] Finally, considering the safety needs of industrial sites, we designed a safety monitoring module7. This module uses a combination of rule-based and anomaly detection methods to monitor key parameters such as pressure, temperature, and vibration in real time. When an anomaly is detected, the system triggers different levels of alarms or protective measures based on the severity of the anomaly to ensure the safety of equipment and personnel.

[0050] By integrating these modules, the present invention achieves intelligent, efficient, and safe air compressor systems. In practical applications, this system has achieved remarkable results across multiple industries. For example, in a large chemical plant, the system saves approximately 2 million kWh of electricity annually, a 16.8% energy saving rate, and reduces carbon emissions by 1,600 tons annually. At a food and beverage manufacturer, the system not only improves production efficiency but also effectively ensures product quality through precise pressure control and cleanliness management.

[0051] In summary, this invention provides a novel solution for the intelligent control of air compressor systems. By incorporating advanced machine learning techniques and integrating them with the practical needs of industrial sites, we have developed a control system that is both intelligent and efficient, as well as safe and reliable. We believe that with continued optimization and widespread application of this technology, this system will make significant contributions to energy conservation, emission reduction, and intelligent upgrades in industrial production.

[0052] It should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, replacements, and improvements made within the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An intelligent control system for air compressor, It is characterized in that include: Data acquisition module, used to obtain the operating parameters of multiple air compressors; a central controller, electrically connected to the data acquisition module, for receiving the operating parameters and generating control instructions based on the operating parameters; A machine learning module, disposed in the central controller, for building a dynamic operation model of the air compressor based on the operating parameters and learning an optimal control strategy; A collaborative control module, electrically connected to the central controller, for determining operating parameters of each air compressor based on the optimal control strategy; an execution module, electrically connected to the collaborative control module, and configured to adjust the working state of each air compressor based on the operating parameters; A human-computer interaction module, electrically connected to the central controller, for displaying the system operation status and receiving user input; as well as The safety monitoring module is electrically connected to the central controller and is used to monitor the safety of system operation and trigger a protection mechanism when an abnormality is detected.

2. According to claim 1, the air compressor intelligent control system, It is characterized in that The data acquisition module comprises: The pressure sensor, flow meter, power analyzer, temperature sensor and vibration sensor are used to obtain the pressure, flow, power, temperature and vibration data of the air compressor respectively.

3. The air compressor intelligent control system according to claim 1, It is characterized in that The central controller comprises: A processor, a memory and a communication interface, wherein the processor is used to execute instructions stored in the memory, and the communication interface is used to exchange data with other modules.

4. The air compressor intelligent control system according to claim 1, It is characterized in that The machine learning module includes: A data preprocessing submodule, used for cleaning, normalizing and feature extraction of the operating parameters; An air compressor model building submodule, used to build a dynamic operation model of each air compressor based on the preprocessed operation parameters; A reinforcement learning submodule, configured to learn an optimal control strategy based on the dynamic operation model; and The demand forecasting submodule is used to forecast the gas demand in the future.

5. The air compressor intelligent control system according to claim 4, It is characterized in that The air compressor model building submodule uses a recurrent neural network to build the dynamic operation model.

6. The air compressor intelligent control system according to claim 4, It is characterized in that The reinforcement learning submodule adopts a deep Q network algorithm to learn the optimal control strategy.

7. The air compressor intelligent control system according to claim 4, It is characterized in that The demand prediction submodule uses a long short-term memory network to predict the gas demand.

8. The air compressor intelligent control system according to claim 1, It is characterized in that The collaborative control module adopts a model predictive control framework to determine the operating parameters of each air compressor based on the optimal control strategy and the gas demand.

9. The air compressor intelligent control system according to claim 1, It is characterized in that Also includes: Edge computing nodes are set up near each air compressor to execute real-time control logic and data preprocessing.

10. The air compressor intelligent control system according to claim 1, It is characterized in that Also includes: The heat recovery module is electrically connected to the central controller and is used to monitor the heat discharged by the air compressor and optimize the operation strategy of the air compressor based on the heat energy utilization efficiency.