Screw air compressor dynamic adjusting method and system based on multi-sensor feedback

Abstract

The invention discloses a screw air compressor dynamic adjustment method and system based on multi-sensor feedback, and relates to the technical field of air compressor adjustment. The method comprises the steps that N air compressor key parts of a target screw air compressor are obtained through division, and a screw air compressor adjusting platform is built according to the N air compressor key parts; n multi-source heterogeneous operation parameters of the N air compressor key parts are collected, state prediction modeling is conducted on the N multi-source heterogeneous operation parameters, and N air compressor state prediction models are generated; feedback regulation analysis is conducted on the N air compressor state prediction models, and target air compressor regulation parameters are determined; and based on matching of the target air compressor adjusting parameters and the adjusting executing mechanism, the adjusting mechanism combination is activated, and air compressor operation dynamic adjusting is executed. The technical problems that in the prior art, a screw type air compressor is insufficient in adjusting precision and lags behind in adjusting response are solved, and the technical effect that accurate dynamic adjusting of the screw type air compressor is achieved based on multi-sensor feedback is achieved.

Description

Technical Field

[0001] This invention relates to the field of air compressor adjustment technology, specifically to a dynamic adjustment method and system for screw air compressors based on multi-sensor feedback. Background Technology

[0002] Screw air compressors are commonly used air source equipment in industrial production, and their operating status directly affects production efficiency and energy consumption. Currently, screw air compressors typically rely on a single or limited set of sensors to collect a small number of operating parameters during operation. Adjustment strategies are often based on static settings or manual experience, resulting in a failure to promptly reflect the actual operating conditions of critical components. This inadequacy in adjustment accuracy and lag becomes particularly pronounced under conditions of large load fluctuations or frequent changes in environmental conditions, easily leading to increased energy consumption, accelerated equipment wear, and decreased operational stability. Summary of the Invention

[0003] This application provides a dynamic adjustment method and system for screw air compressors based on multi-sensor feedback, which solves the technical problems of insufficient adjustment accuracy and lag in adjustment response of screw air compressors in the prior art.

[0004] The first aspect of this application provides a dynamic adjustment method for a screw air compressor based on multi-sensor feedback, the method comprising: The target screw air compressor is divided into N key components. Based on these N key components, a screw air compressor adjustment platform is constructed, comprising a multi-parameter sensor network, an edge computing unit, and adjustment actuators. The multi-parameter sensor network collects N multi-source heterogeneous operating parameters from the N key components. The edge computing unit performs state prediction modeling on these N parameters to generate N air compressor state prediction models. Based on multiple air compressor adjustment objectives, an air compressor objective function is designed. This objective function is used to perform feedback adjustment analysis on the N air compressor state prediction models to determine the target air compressor adjustment parameters. Based on the matching of the target air compressor adjustment parameters with the adjustment actuators, an adjustment mechanism combination is activated, and the adjustment mechanism combination performs dynamic adjustment of the air compressor operation according to the target air compressor adjustment parameters.

[0005] A second aspect of this application provides a dynamic adjustment system for a screw air compressor based on multi-sensor feedback, the system comprising: Platform Construction Module: The target screw air compressor is divided into N key components. Based on these N key components, a screw air compressor adjustment platform is built, including a multi-parameter sensor network, an edge computing unit, and adjustment actuators. Modeling Module: The multi-parameter sensor network collects N multi-source heterogeneous operating parameters from the N key components. The edge computing unit performs state prediction modeling on these N parameters, generating N air compressor state prediction models. Analysis Module: Based on the multiple objectives of air compressor adjustment, an air compressor objective function is designed. This objective function is used to perform feedback adjustment analysis on the N air compressor state prediction models to determine the target air compressor adjustment parameters. Dynamic Adjustment Module: Based on the matching of the target air compressor adjustment parameters with the adjustment actuators, a combination of adjustment mechanisms is activated. This combination then performs dynamic adjustment of the air compressor operation according to the target air compressor adjustment parameters.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, N key components of the target screw air compressor are identified. Based on these N key components, a screw air compressor adjustment platform is constructed, comprising a multi-parameter sensor network, an edge computing unit, and adjustment actuators. Next, N multi-source heterogeneous operating parameters from the N key components are collected via the multi-parameter sensor network. The edge computing unit then performs state prediction modeling on these N parameters, generating N air compressor state prediction models. Then, based on the multiple objectives of air compressor adjustment, an air compressor objective function is designed. This objective function is used to perform feedback adjustment analysis on the N air compressor state prediction models to determine the target air compressor adjustment parameters. Finally, based on the matching of the target air compressor adjustment parameters with the adjustment actuators, the adjustment mechanism combination is activated, and the air compressor's dynamic operation is dynamically adjusted according to the target air compressor adjustment parameters. This solves the technical problems of insufficient adjustment accuracy and lag in the adjustment response of screw air compressors in existing technologies, achieving the technical effect of precise dynamic adjustment of screw air compressors based on multi-sensor feedback. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 A schematic flowchart of a dynamic adjustment method for a screw air compressor based on multi-sensor feedback provided in an embodiment of this application; Figure 2 This is a schematic diagram of the dynamic adjustment system for a screw air compressor based on multi-sensor feedback, provided in an embodiment of this application.

[0009] Figure labeling: Platform building module 11, modeling module 12, analysis module 13, dynamic adjustment module 14. Detailed Implementation

[0010] This application provides a method and system for dynamic adjustment of screw air compressors based on multi-sensor feedback, which solves the technical problems of insufficient adjustment accuracy and sluggish adjustment response in the prior art of screw air compressors.

[0011] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0012] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0013] Example 1, as Figure 1 As shown, this application provides a dynamic adjustment method for a screw air compressor based on multi-sensor feedback, wherein the method includes: The target screw air compressor is divided into N key parts. Based on the N key parts, a screw air compressor adjustment platform is built. The screw air compressor adjustment platform includes a multi-parameter sensor network, an edge computing unit, and an adjustment actuator.

[0014] In this embodiment of the application, for the target screw air compressor, based on its structural attributes and operating functions, and combined with finite element simulation and / or on-site working condition simulation and historical operating data, the unit is quantitatively evaluated according to its sensitivity to key performance indicators (such as outlet pressure, flow rate, compression efficiency, temperature rise of various parts, vibration amplitude, leakage tendency, oil temperature / oil quality, electric drive current, etc.) and fault impact, thereby determining several candidate key parts; the candidate parts are sorted from high to low according to their sensitivity scores, and the top N parts with the highest scores are selected as N key parts of the air compressor.

[0015] Based on N key components of an air compressor, a multi-parameter sensor network is constructed: several types of sensors are matched for each key component (e.g., temperature sensors, pressure sensors, flow sensors, acceleration / vibration sensors, speed / encoders, acoustic / ultrasonic sensors, oil monitoring sensors, current / voltage sensors, displacement / strain sensors, or infrared thermal imaging modules, etc.), and the deployment points, installation methods, sampling frequencies, range and accuracy requirements, calibration methods, and protection levels are determined for each type of sensor; a hybrid topology is adopted according to the requirements of real-time performance, bandwidth, and cabling feasibility (wired industrial bus / Ethernet or Time-Sensitive Networking is preferred for key real-time quantities, while wireless links are used for quantities that are difficult to cable or low-frequency quantities), and timestamp synchronization (e.g., PTP / NTP) and local preliminary filtering, noise reduction, anomaly removal, and data compression are implemented at the sensor end to reduce communication load.

[0016] Edge computing units are selected and configured according to the data processing and communication requirements of each key component. The edge computing units are preferably deployed in the field control cabinet or the nearest chassis of the unit, and have a data acquisition gateway, real-time data buffer, local preprocessing (filtering, framing, feature extraction such as time domain RMS, kurtosis, peak value, frequency domain spectrum analysis, etc.), model inference engine, online model update module, historical data storage and remote communication interface. Its hardware performance (CPU / GPU / FPGA, storage, number of I / O) and software stack (real-time operating system or containerized inference environment, support for OPC-UA / MQTT / Modbus protocols, security authentication and encryption) are determined according to the maximum latency constraints and the number of concurrent models.

[0017] The selection of the regulating actuator is based on the physical realization of the target regulating variable, including but not limited to frequency converters (VFDs) for motor speed regulation, electromagnetic or servo valves for intake / exhaust volume and bypass control, throttling devices, variable frequency control of fuel injection volume / cooling fan and cooling water pump, mechanical regulating mechanisms (such as bearing preload mechanisms, spool valve positioners), etc.; and the control interface (4–20mA / 0–10V analog, digital IO, Modbus / EtherCAT, etc.), action delay, control accuracy and redundancy scheme of each actuator are specified. In terms of communication integration, the multi-parameter sensor network, edge computing unit, and regulating actuator are networked through a unified communication layer (supporting industrial protocols and message buses). The edge unit undertakes local closed-loop fast control and safety limiting (local PID / nested loop), while reporting feature and state prediction results to the upper control / optimization layer to execute multi-objective optimization strategies. The system design includes fault detection and degradation strategies (1002 redundancy for key sensors, safety bypass and backoff conditions when sensors / actuators are abnormal), timing synchronization and time consistency guarantees, as well as on-site installation, grounding, shielding, and protection measures.

[0018] Furthermore, the target screw air compressor is divided into N key components, including: Finite element simulation is performed based on the structural properties and operational functions of the target screw air compressor to obtain an air compressor operation simulation model. A working condition simulation parameter table is constructed based on the application information of the target screw air compressor. The working condition simulation parameter table is applied to the air compressor operation simulation model for operation simulation recording to obtain air compressor operation response data. Based on the air compressor operation response data, key components of the target screw air compressor are divided to obtain the N key components of the air compressor.

[0019] Based on the structural attributes of the target screw air compressor, including its 3D structural model, component material properties, assembly relationships, and transmission methods, as well as operational functional information such as gas compression process, lubrication and cooling paths, and load variation characteristics, a finite element simulation model is established using finite element analysis software. This model includes major components such as the casing, rotor, bearings, seals, and cooling system. Physical parameters such as the elastic modulus, coefficient of thermal expansion, and coefficient of friction of the actual materials are incorporated into the model, and mesh elements are divided to meet accuracy requirements, ensuring the accuracy and computational efficiency of the simulation results. According to the application scenarios and design specifications of the target screw air compressor, a working condition simulation parameter table covering common and extreme operating conditions is constructed. This table includes, but is not limited to, ambient temperature and humidity, intake pressure, exhaust pressure, load variation amplitude and rate, lubricating oil temperature, cooling water flow rate, motor speed, and working cycle. Specific load spectra and disturbance modes can be customized for different application industries. The operating condition simulation parameters were applied item by item to the air compressor operation simulation model, and multiple rounds of operation simulations were performed. Multi-dimensional operational response data, including stress-strain distribution, heat distribution, vibration response, airflow pressure changes, and efficiency changes, were recorded for each component under steady-state and transient conditions. Statistical and sensitivity analyses were performed on the obtained air compressor operational response data to assess the impact of each component on the overall machine performance, energy consumption, and failure risk under different operating conditions. Principal component analysis and modal analysis were used to identify components with large response amplitudes, high frequency of change, high sensitivity to performance impact, and obvious early signs of failure. Finally, the top N components, ranked by their degree of impact, were selected as the critical components of the air compressor.

[0020] Furthermore, the construction of a screw air compressor adjustment platform includes: A monitoring requirement analysis and sensor deployment analysis are performed on the N key components of the air compressor to obtain a set of sensor deployment parameters for each of the N components. A multi-parameter sensor network is obtained by deploying sensors on the N key components according to these parameters. An edge computing unit is selected and configured based on the data processing and communication requirements of the N key components. An adjustment actuator is obtained based on the structural properties of the target screw air compressor. The multi-parameter sensor network, the edge computing unit, and the adjustment actuator are then integrated to establish a screw air compressor adjustment platform.

[0021] For N key components of an air compressor, a monitoring requirements analysis is conducted based on their working characteristics, stress conditions, temperature rise distribution, fluid state changes, and possible failure modes. This determines the types of physical quantities (such as temperature, pressure, flow rate, vibration, noise, oil quality, current, voltage, etc.) that need to be collected for each key component, as well as the sampling accuracy, response time, and environmental adaptability requirements. Simultaneously, considering the structural layout of the components and the available installation space, a sensor deployment analysis is performed to determine the sensor type, installation location, installation method, protection level, power supply method, and signal output interface, thus forming a set of sensor deployment parameters for N components. Based on the parameter set of N sensors, sensor topology is deployed for N key parts of the air compressor. A hybrid wired and wireless networking method is preferred. Industrial Ethernet, CAN bus or TSN network is used for high-speed and real-time acquisition channels, while wireless communication methods such as ZigBee, LoRa or Wi-Fi can be used for low-speed and non-critical acquisition channels. Time consistency of multi-source acquisition data is ensured through time synchronization protocols (PTP / NTP), and preliminary filtering, anomaly detection and data compression functions are configured at the sensor end or near-end acquisition nodes to form a multi-parameter sensor network covering all key parts.

[0022] Based on the data volume and latency requirements of N key components of the air compressor in terms of operation status prediction and adjustment decision calculation, a suitable edge computing unit is selected and configured. The edge computing unit preferably has functions such as multi-channel high-speed data acquisition interface, real-time operating system, parallel computing capability (CPU / GPU / FPGA), local data caching, model inference acceleration, communication protocol conversion, and security encryption, and can communicate with the upper control system through Ethernet, industrial bus or wireless gateway.

[0023] Based on the structural properties and adjustment requirements of the target screw air compressor, select appropriate adjustment actuators, including but not limited to motor frequency conversion control unit, intake regulating valve, exhaust back pressure regulating valve, oil circuit regulating valve, cooling water pump / fan frequency converter, slide valve positioner, lubricating oil injection control unit, etc.; at the same time, specify the control interface type, control accuracy, action delay and safety protection strategy of each actuator.

[0024] By integrating multi-parameter sensor networks, edge computing units, and regulating actuators through a unified communication architecture, a closed-loop path of data acquisition, state prediction, regulation decision-making, and execution feedback is established, enabling the construction and operation of a screw air compressor regulation platform.

[0025] The multi-parameter sensor network collects N multi-source heterogeneous operating parameters of the N key parts of the air compressor, and the edge computing unit performs state prediction modeling on the N multi-source heterogeneous operating parameters to generate N air compressor state prediction models.

[0026] Through an established multi-parameter sensor network, real-time operational data is collected from N key components of the air compressor. This operational data comprises multi-source heterogeneous parameters, including but not limited to temperature, pressure, flow rate, vibration acceleration, rotational speed, power, current, voltage, oil particle size, lubricating oil viscosity, noise spectrum characteristics, shaft displacement, and cooling water temperature. These N multi-source heterogeneous operational parameters from the N key components of the air compressor are input into an edge computing unit for state prediction modeling, generating N air compressor state prediction models.

[0027] Furthermore, N air compressor state prediction models are generated, including: Based on the edge computing unit, a preprocessing node and a prediction model generation node are determined; the preprocessing node performs outlier cleaning on the N multi-source heterogeneous operating parameters according to data application standards to obtain N usable multi-source heterogeneous operating parameters; the N usable multi-source heterogeneous operating parameters are standardized to obtain N standard key component operating parameters; based on the prediction model generation node, the N standard key component operating parameters are used for state prediction modeling to generate N air compressor state prediction models.

[0028] First, based on the hardware architecture and computing resource allocation of the edge computing unit, preprocessing nodes for data preprocessing and prediction model generation nodes for model training and inference are determined. These nodes can be independent processing cores, dedicated computing modules (such as GPUs / FPGAs), or software threads with specific functions within the edge computing unit. Before the data enters the prediction modeling process, the preprocessing nodes perform data quality checks on N multi-source heterogeneous operating parameters. The checks include sampling timestamp consistency, sampling accuracy compliance, data amplitude out-of-bounds detection, signal-to-noise ratio analysis, and sampling interval stability. Outlier cleaning is performed on the detected abnormal data. Sliding window statistics are used to remove outliers that significantly deviate from the mean. For missing values ​​caused by communication jitter or instantaneous sensor failure, nearest neighbor interpolation, historical mean compensation, or multivariate regression compensation based on relevant variables are used to obtain N usable multi-source heterogeneous operating parameters. Subsequently, the N available multi-source heterogeneous operating parameters are standardized according to a unified data application standard, including but not limited to unit unification (e.g., pressure is unified to MPa, temperature to ℃), numerical normalization (e.g., conversion to interval [0,1] or standard normal distribution), and time-aligned resampling (ensuring that multi-sensor data are aligned at the same sampling time). This yields N standard key component operating parameters, ensuring the consistency and comparability of the data input into the prediction model. Finally, the N standard key component operating parameters are input to the prediction model generation node. Based on the dynamic characteristics and state evolution patterns of different key components, suitable modeling algorithms are selected or combined. These include statistical learning-based Autoregressive Moving Average (ARMA) and ARIMA models, machine learning-based Support Vector Regression (SVR) and Random Forest Regression (RFR), and deep learning-based Long Short-Term Memory Network (LSTM), Gated Recurrent Unit (GRU), and Convolutional Neural Network (CNN) time series modeling. The prediction accuracy of the model at different operating stages can be improved through transfer learning or online incremental training. After modeling is completed, state prediction models corresponding to N key parts of the air compressor are generated. Each model can output the trend prediction of the future operating state of the corresponding part and the potential risk assessment results under real-time input conditions, which can be used to drive subsequent dynamic adjustment decisions.

[0029] Furthermore, based on the prediction model generation nodes, state prediction modeling is performed on the operating parameters of the N standard key components to generate N air compressor state prediction models, including: Based on the N key components of the air compressor, N key component prediction targets are determined; based on the prediction targets of the N key components and the data characteristic information of the operating parameters of the N standard key components, a prediction network structure for the N key components is selected; through the prediction model generation node, the N key component prediction network structure is used to perform state prediction modeling on the operating parameters of the N standard key components, generating N key component state prediction models; an incremental learning mechanism is introduced to update the N key component state prediction models online, generating the N air compressor state prediction models.

[0030] First, based on the functional characteristics and operational requirements of N key components of the air compressor, prediction targets for each key component are determined, such as temperature change trends, pressure fluctuation ranges, vibration anomaly indicators, speed stability, and current load changes, forming a set of prediction targets for N key components. Second, combining the prediction targets of the N key components with the data characteristics of the corresponding standard key component operating parameters, such as the fluctuation frequency of the time series, data noise distribution, correlation between variables, and historical fault modes, the most suitable prediction network structure is selected for each key component. These network structures include, but are not limited to: Long Short-Term Memory (LSTM) networks, Gated Recurrent Units (GRUs), Convolutional Neural Networks (CNNs), Autoregressive Neural Networks (RNNs), and Temporal Convolutional Networks (TCNs). By comparing the modeling capabilities and prediction accuracy of the networks, a dedicated prediction network structure for each key component is determined. Subsequently, using the prediction model generation nodes and the selected key component prediction network structures, the state prediction modeling process is executed for the standard operating parameter input data of the N key components, generating state prediction models for each of the N key components. Each prediction model can output the operational status trend and anomaly warning indicators for the corresponding key component within a certain future time window. Furthermore, the system introduces an incremental learning mechanism to update the state prediction models for the aforementioned N key components online. This incremental learning mechanism allows the prediction models to dynamically adjust and optimize their parameters through small-batch incremental training or transfer learning algorithms as the edge computing unit continuously receives newly collected operational data. This adapts to changes in the air compressor's operating environment, equipment wear, and external disturbances, improving the model's adaptability and prediction accuracy. Ultimately, the updated state prediction models for the N key components constitute the N air compressor state prediction models, which are then used by the subsequent adjustment and decision-making module.

[0031] Based on the multiple objectives of air compressor regulation, an air compressor objective function is designed. The air compressor objective function is then used to perform feedback regulation analysis on the N air compressor state prediction models to determine the target air compressor regulation parameters.

[0032] Air compressor adjustment targets include energy efficiency indicators (such as unit compression power and energy saving rate), operational stability indicators (such as vibration amplitude and temperature fluctuation range), equipment life extension indicators (such as stress distribution and wear rate of key components), safety indicators (such as overpressure and overheat protection), and gas production quality indicators (such as pressure stability and flow accuracy).

[0033] Based on the multi-objective regulation of air compressors, a comprehensive objective function for air compressors is designed. Each indicator is weighted and normalized through weighting coefficients to form a unified mathematical expression, which facilitates quantitative evaluation of the regulation effect.

[0034] A feedback adjustment analysis is performed on N air compressor state prediction models generated by edge computing units using an air compressor objective function. Specifically, by inputting the future operating state parameters output by the prediction models into the objective function, the achievement degree and existing deviations of each objective under the current operating state are evaluated. Combining historical operating data and expert system rules, a multi-objective optimization algorithm (such as genetic algorithm, particle swarm optimization, multi-objective ant colony optimization, etc.) is used to iteratively search within the air compressor adjustment parameter space to find the optimal combination of adjustment parameters that maximizes the objective function value. Finally, the target air compressor adjustment parameters that meet the requirements of multi-objective optimization are determined, including but not limited to motor speed setpoint, intake and exhaust valve opening, cooling system flow rate adjustment, and lubricating oil injection rate, providing precise adjustment commands to the actuators and achieving dynamic optimization operation of the screw air compressor.

[0035] Furthermore, the objective function for designing the air compressor includes: A correlation index analysis is performed on each objective in the multi-objective regulation of the air compressor to obtain a set of correlation indexes for the multi-objective regulation. Principal component analysis and index screening are then performed on the set of correlation indexes for the multi-objective regulation to obtain a set of key indicators for the multi-objective regulation. An evaluation function is designed and fitted on the set of key indicators for the multi-objective regulation using an air compressor expert system and historical air compressor operating data to generate an evaluation function for the multi-objective regulation. The evaluation function for the multi-objective regulation is then weighted and normalized to determine the objective function of the air compressor.

[0036] Air compressor regulation typically encompasses multiple objectives, including energy efficiency optimization (e.g., unit energy consumption, compression efficiency), operational stability (e.g., vibration amplitude, temperature fluctuation range), equipment reliability (e.g., stress and wear of key components), safety assurance (e.g., pressure and temperature over-limit protection), and gas production quality (e.g., pressure and flow stability). By collecting multidimensional monitoring data and historical operating records related to these objectives, a set of correlation indicators for these regulation objectives is constructed. For the obtained set of correlation indicators, principal component analysis (PCA) is used to reduce the dimensionality of the indicators, eliminating those with low contributions. Combined with indicator screening techniques (e.g., threshold screening based on contribution rate, Lasso regression, etc.), a set of key indicators representing the multi-objective regulation needs is selected, ensuring the simplicity and efficiency of the objective function design.

[0037] Based on the empirical rules and historical operating data of the air compressor expert system, an evaluation function is designed and fitted for a set of key indicators for multiple adjustment objectives. Specifically, statistical regression analysis, machine learning regression models, or fuzzy logic systems are used to establish the mapping relationship between each key indicator and the adjustment objective, forming a multi-adjustment objective evaluation function. This multi-adjustment objective evaluation function can comprehensively evaluate the input key indicator data, reflecting the performance of the system under the current adjustment strategy, taking into account the weights and mutual constraints of different indicators, and achieving a comprehensive evaluation of multiple objectives. Finally, the multi-adjustment objective evaluation function is weighted according to preset weight coefficients, and normalization methods (such as Min-Max normalization and Z-score standardization) are applied to unify the dimensions and scale, resulting in the comprehensive air compressor objective function.

[0038] Furthermore, determining the target air compressor adjustment parameters includes: A correlation influence analysis is performed on the N key components of the air compressor to generate a cascaded influence network of the key components' operation. Based on the cascaded influence network of the key components' operation, a feedback adjustment analysis is performed on the N air compressor state prediction models to obtain the air compressor adjustment parameter space. The air compressor objective function is used to evaluate cross-variation and iteratively solve for optimization within the air compressor adjustment parameter space to determine the target air compressor adjustment parameters.

[0039] First, an in-depth analysis of the operational correlations among N key components of the air compressor is conducted. By collecting and mining operational status data of each key component, and combining structural mechanics and fluid mechanics models, an operational cascade influence network among the key components is constructed. The operational cascade influence network uses each key component as a node, with edges between nodes representing causal or dependent relationships in operational states, and weights are assigned to reflect the intensity of influence and the transmission path. Subsequently, based on the operational cascade influence network of key components, this network is embedded into N air compressor state prediction models to achieve joint feedback regulation analysis of the coupled states of multiple components. Through this joint analysis, an air compressor regulation parameter space containing state variables and mutually influencing factors of each key component can be obtained, clarifying the coupling relationships and constraints between parameters. Finally, using the air compressor objective function, a multi-objective optimization algorithm is employed within this regulation parameter space for evaluation, crossover, mutation, and iterative solution. The combination of regulation parameters can be continuously optimized through genetic algorithms, particle swarm optimization, or other evolutionary computation methods, with the goal of maximizing the performance index of the objective function to ensure that the air compressor achieves its optimal operating state under multiple objectives. After the iteration terminates, the target air compressor adjustment parameters that meet the adjustment requirements and are globally optimal or near optimal are determined, and the actuators are provided with these parameters for dynamic adjustment.

[0040] Furthermore, the generation of a cascading impact network for critical components includes: Each of the N key components of the air compressor is used as node information for control correlation analysis to obtain a key component correlation node network; the influence degree of each correlation node in the key component correlation node network is quantified by the analytic hierarchy process to determine the influence measurement factor of the correlation node; based on the influence measurement factor of the correlation node, the key component correlation node network is labeled with directed edges to generate the key component operation cascade influence network.

[0041] By collecting real-time and historical operational data from key components, and combining this data with equipment structural characteristics and operating mechanisms, statistical methods (such as correlation coefficient analysis and mutual information analysis) and control system theory (such as causal inference and Granger causality tests) are used to identify the mutual influence relationships between key components. By analyzing the synchronicity of state changes, response delays, and feedback paths between nodes, a preliminary network of key component-related nodes is constructed to determine which key components have direct or indirect control or influence on other components. Secondly, the Analytic Hierarchy Process (AHP) is used to quantitatively assess the influence degree of each related node in the key component-related node network. Through expert scoring, historical operational data feedback, and model simulation results, a judgment matrix is ​​constructed, and the influence weights between nodes are calculated to obtain quantitative influence measurement factors for related nodes, reflecting the interaction strength and transmission effect between key components. Finally, based on the aforementioned influence measurement factors for related nodes, directed edges are labeled on the edges of the key component-related node network to clarify the direction and weight of influence, generating a complete cascaded influence network of key component operations. The cascading impact network of critical components not only reflects the dependencies between critical components, but also reflects the path and intensity of changes in operating status that may propagate along the network, providing a basis for subsequent construction of adjustment parameter space and multi-objective feedback adjustment.

[0042] Furthermore, obtaining the air compressor adjustment parameter space includes: The critical component operation cascaded influence network is embedded into the N air compressor state prediction models for integrated prediction coupling to obtain a multi-component coupled state prediction model; based on the output results of the N air compressor state prediction models, the component abnormal deviation parameter set is determined; based on the multi-component coupled state prediction model, the adjustment parameter range of the component abnormal deviation parameter set is analyzed to obtain the air compressor adjustment parameter space.

[0043] First, the cascading influence network of key components is embedded into N air compressor state prediction models to construct a multi-component coupled state prediction model. Second, based on the outputs of the N air compressor state prediction models, the predicted states of each key component are analyzed to identify abnormal deviation parameter sets, i.e., key state parameters deviating from the normal operating range or design standards, such as excessive temperature, abnormal vibration, and excessive pressure fluctuations. Subsequently, based on the multi-component coupled state prediction model, the adjustment parameter ranges for the abnormal deviation parameter sets are analyzed. Specifically, considering equipment operating condition limitations, controller adjustable range, and safety constraints, reasonable variation ranges for each adjustment parameter (such as valve opening, speed setting, and current adjustment) are derived, forming a dynamic air compressor adjustment parameter space. This air compressor adjustment parameter space effectively covers all possible adjustment schemes to ensure the safe and stable operation of the air compressor.

[0044] Based on the matching of the target air compressor adjustment parameters with the adjustment actuator, the adjustment mechanism combination is activated, and the air compressor operation is dynamically adjusted according to the target air compressor adjustment parameters through the adjustment mechanism combination.

[0045] Based on the target air compressor adjustment parameters, the system matches this parameter set with pre-set adjustment actuators to ensure that the adjustment parameters correspond to the control range, response characteristics, and action accuracy of each actuator. The matching process includes parameter format conversion, control signal adjustment, and actuator status confirmation. Subsequently, based on the matching results, the corresponding adjustment mechanism combination is activated. These adjustment mechanisms may include intake regulating valves, exhaust back pressure valves, motor frequency converter control units, cooling system pumps and valves, lubrication system injection devices, etc. Under the coordinated control of the edge computing unit, the adjustment mechanism combination performs dynamic adjustment operations according to the target air compressor adjustment parameters. Specifically, the system sends control commands in real time to cause each actuator to adjust its operating status according to predetermined parameters, such as adjusting valve opening, adjusting speed settings, current magnitude, and lubricating oil flow, to ensure that the air compressor achieves maximum energy efficiency, stable operation, and safety assurance under multi-objective optimization.

[0046] In summary, the embodiments of this application have at least the following technical effects: First, N key components of the target screw air compressor are identified. Based on these N key components, a screw air compressor adjustment platform is constructed, comprising a multi-parameter sensor network, an edge computing unit, and adjustment actuators. Next, N multi-source heterogeneous operating parameters from the N key components are collected via the multi-parameter sensor network. The edge computing unit then performs state prediction modeling on these N parameters, generating N air compressor state prediction models. Then, based on the multiple objectives of air compressor adjustment, an air compressor objective function is designed. This objective function is used to perform feedback adjustment analysis on the N air compressor state prediction models to determine the target air compressor adjustment parameters. Finally, based on the matching of the target air compressor adjustment parameters with the adjustment actuators, the adjustment mechanism combination is activated, and the air compressor's dynamic operation is dynamically adjusted according to the target air compressor adjustment parameters. This solves the technical problems of insufficient adjustment accuracy and lag in the adjustment response of screw air compressors in existing technologies, achieving the technical effect of precise dynamic adjustment of screw air compressors based on multi-sensor feedback.

[0047] Example 2 is based on the same inventive concept as the multi-sensor feedback-based dynamic adjustment method for screw air compressors in the previous examples, such as... Figure 2 As shown, this application provides a dynamic adjustment system for a screw air compressor based on multi-sensor feedback, wherein the system includes: Platform Construction Module 11: N key components of the target screw air compressor are identified. Based on these N key components, a screw air compressor adjustment platform is built, comprising a multi-parameter sensor network, an edge computing unit, and an adjustment actuator. Modeling Module 12: N multi-source heterogeneous operating parameters of the N key components are collected through the multi-parameter sensor network. Based on the edge computing unit, state prediction modeling is performed on these N multi-source heterogeneous operating parameters to generate N air compressor state prediction models. Analysis Module 13: Based on the multiple objectives of air compressor adjustment, an air compressor objective function is designed. The air compressor objective function is used to perform feedback adjustment analysis on the N air compressor state prediction models to determine the target air compressor adjustment parameters. Dynamic Adjustment Module 14: Based on the matching of the target air compressor adjustment parameters with the adjustment actuator, an adjustment mechanism combination is activated, and the air compressor operation is dynamically adjusted according to the target air compressor adjustment parameters through the adjustment mechanism combination.

[0048] Furthermore, the platform building module 11 is used to perform the following methods: Finite element simulation is performed based on the structural properties and operational functions of the target screw air compressor to obtain an air compressor operation simulation model. A working condition simulation parameter table is constructed based on the application information of the target screw air compressor. The working condition simulation parameter table is applied to the air compressor operation simulation model for operation simulation recording to obtain air compressor operation response data. Based on the air compressor operation response data, key components of the target screw air compressor are divided to obtain the N key components of the air compressor.

[0049] Furthermore, the platform building module 11 is used to perform the following methods: A monitoring requirement analysis and sensor deployment analysis are performed on the N key components of the air compressor to obtain a set of sensor deployment parameters for each of the N components. A multi-parameter sensor network is obtained by deploying sensors on the N key components according to these parameters. An edge computing unit is selected and configured based on the data processing and communication requirements of the N key components. An adjustment actuator is obtained based on the structural properties of the target screw air compressor. The multi-parameter sensor network, the edge computing unit, and the adjustment actuator are then integrated to establish a screw air compressor adjustment platform.

[0050] Furthermore, the modeling module 12 is used to perform the following methods: Based on the edge computing unit, a preprocessing node and a prediction model generation node are determined; the preprocessing node performs outlier cleaning on the N multi-source heterogeneous operating parameters according to data application standards to obtain N usable multi-source heterogeneous operating parameters; the N usable multi-source heterogeneous operating parameters are standardized to obtain N standard key component operating parameters; based on the prediction model generation node, the N standard key component operating parameters are used for state prediction modeling to generate N air compressor state prediction models.

[0051] Furthermore, the modeling module 12 is used to perform the following methods: Based on the N key components of the air compressor, N key component prediction targets are determined; based on the prediction targets of the N key components and the data characteristic information of the operating parameters of the N standard key components, a prediction network structure for the N key components is selected; through the prediction model generation node, the N key component prediction network structure is used to perform state prediction modeling on the operating parameters of the N standard key components, generating N key component state prediction models; an incremental learning mechanism is introduced to update the N key component state prediction models online, generating the N air compressor state prediction models.

[0052] Furthermore, the analysis module 13 is used to perform the following methods: A correlation index analysis is performed on each objective in the multi-objective regulation of the air compressor to obtain a set of correlation indexes for the multi-objective regulation. Principal component analysis and index screening are then performed on the set of correlation indexes for the multi-objective regulation to obtain a set of key indicators for the multi-objective regulation. An evaluation function is designed and fitted on the set of key indicators for the multi-objective regulation using an air compressor expert system and historical air compressor operating data to generate an evaluation function for the multi-objective regulation. The evaluation function for the multi-objective regulation is then weighted and normalized to determine the objective function of the air compressor.

[0053] Furthermore, the analysis module 13 is used to perform the following methods: A correlation influence analysis is performed on the N key components of the air compressor to generate a cascaded influence network of the key components' operation. Based on the cascaded influence network of the key components' operation, a feedback adjustment analysis is performed on the N air compressor state prediction models to obtain the air compressor adjustment parameter space. The air compressor objective function is used to evaluate cross-variation and iteratively solve for optimization within the air compressor adjustment parameter space to determine the target air compressor adjustment parameters.

[0054] Furthermore, the analysis module 13 is used to perform the following methods: Each of the N key components of the air compressor is used as node information for control correlation analysis to obtain a key component correlation node network; the influence degree of each correlation node in the key component correlation node network is quantified by the analytic hierarchy process to determine the influence measurement factor of the correlation node; based on the influence measurement factor of the correlation node, the key component correlation node network is labeled with directed edges to generate the key component operation cascade influence network.

[0055] Furthermore, the analysis module 13 is used to perform the following methods: The critical component operation cascaded influence network is embedded into the N air compressor state prediction models for integrated prediction coupling to obtain a multi-component coupled state prediction model; based on the output results of the N air compressor state prediction models, the component abnormal deviation parameter set is determined; based on the multi-component coupled state prediction model, the adjustment parameter range of the component abnormal deviation parameter set is analyzed to obtain the air compressor adjustment parameter space.

[0056] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0057] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0058] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A dynamic adjustment method for screw air compressors based on multi-sensor feedback, characterized in that, The method includes: The target screw air compressor is divided into N key parts. Based on the N key parts, a screw air compressor adjustment platform is built. The screw air compressor adjustment platform includes a multi-parameter sensor network, an edge computing unit, and an adjustment actuator. The multi-parameter sensor network collects N multi-source heterogeneous operating parameters of the N key parts of the air compressor, and the edge computing unit performs state prediction modeling on the N multi-source heterogeneous operating parameters to generate N air compressor state prediction models. Based on the multiple objectives of air compressor regulation, an air compressor objective function is designed, and the air compressor objective function is used to perform feedback regulation analysis on the N air compressor state prediction models to determine the target air compressor regulation parameters. Based on the matching of the target air compressor adjustment parameters with the adjustment actuator, the adjustment mechanism combination is activated, and the air compressor operation is dynamically adjusted according to the target air compressor adjustment parameters through the adjustment mechanism combination.

2. The dynamic adjustment method for screw air compressors based on multi-sensor feedback as described in claim 1, characterized in that, The division yields N key components of the target screw air compressor, including: Finite element simulation is performed on the structural properties and operating functions of the target screw air compressor to obtain an air compressor operation simulation model. Based on the application information of the target screw air compressor, construct a working condition simulation parameter table; The operating condition simulation parameter table is applied to the air compressor operation simulation model to perform operation simulation recording and obtain air compressor operation response data. Based on the air compressor operation response data, the target screw air compressor is divided into key parts to obtain the N key parts of the air compressor.

3. The dynamic adjustment method for screw air compressors based on multi-sensor feedback as described in claim 2, characterized in that, The aforementioned screw air compressor adjustment platform includes: The monitoring requirements and sensor layout analysis of the N key parts of the air compressor are performed to obtain the sensor layout parameter set of the N parts. According to the sensor deployment parameter set of the N parts, the sensor topology is deployed for the N key parts of the air compressor to obtain a multi-parameter sensor network. Based on the data processing and data communication requirements of the N key components of the air compressor, select and configure edge computing units; Based on the structural properties of the target screw air compressor, an adjustment actuator is obtained, and the multi-parameter sensor network, the edge computing unit, and the adjustment actuator are integrated for communication to build the screw air compressor adjustment platform.

4. The dynamic adjustment method for screw air compressors based on multi-sensor feedback as described in claim 1, characterized in that, The generation of N air compressor state prediction models includes: Based on the edge computing unit, the preprocessing node and the prediction model generation node are determined; The preprocessing node performs outlier cleaning on the N multi-source heterogeneous operating parameters according to the data application standard to obtain N usable multi-source heterogeneous operating parameters. The N available multi-source heterogeneous operating parameters are standardized to obtain N standard key component operating parameters; Based on the prediction model, the operating parameters of the N standard key components are generated to perform state prediction modeling, thus generating N air compressor state prediction models.

5. The dynamic adjustment method for a screw air compressor based on multi-sensor feedback as described in claim 4, characterized in that, The process of generating nodes based on the prediction model to perform state prediction modeling on the operating parameters of the N standard key components, generating N air compressor state prediction models, includes: Based on the N key components of the air compressor, determine the predicted targets for the N key components; Based on the data characteristics of the predicted targets of the N key parts and the operating parameters of the N standard key parts, select the prediction network structure for the N key parts; The prediction model generation node uses the N key component prediction network structure to perform state prediction modeling on the operating parameters of the N standard key components, thereby generating N key component state prediction models. An incremental learning mechanism is introduced to update the state prediction models of the N key parts online, thereby generating the N air compressor state prediction models.

6. The dynamic adjustment method for a screw air compressor based on multi-sensor feedback as described in claim 1, characterized in that, The objective function for designing the air compressor includes: The correlation index analysis of each objective in the multi-objective regulation of the air compressor is performed to obtain the multi-objective correlation index set; Principal component analysis and index screening were performed on the set of correlation indicators of the multiple regulation targets to obtain the set of key indicators of the multiple regulation targets. The evaluation function of the set of key indicators of the multi-adjustment target is designed and fitted sequentially using the air compressor expert system and historical air compressor operation data to generate the multi-adjustment target evaluation function. The multi-adjustment objective evaluation function is weighted and normalized to determine the objective function of the air compressor.

7. The dynamic adjustment method for a screw air compressor based on multi-sensor feedback as described in claim 1, characterized in that, The determination of the target air compressor adjustment parameters includes: A correlation impact analysis was performed on the N key components of the air compressor to generate a cascaded impact network of the key components' operation. Based on the cascaded influence network of the key components, feedback adjustment analysis is performed on the N air compressor state prediction models to obtain the air compressor adjustment parameter space. The objective function of the air compressor is used to evaluate cross-variation and iteratively solve for optimization within the air compressor adjustment parameter space to determine the target air compressor adjustment parameters.

8. The dynamic adjustment method for a screw air compressor based on multi-sensor feedback as described in claim 7, characterized in that, The generation of the critical component operation cascaded influence network includes: By taking each of the N key parts of the air compressor as node information, control correlation analysis is performed to obtain the key part correlation node network. The influence degree of each associated node in the network of associated nodes of the key parts is quantified by the analytic hierarchy process, and the influence measurement factor of the associated nodes is determined. Based on the influence metric factor of the associated nodes, the directed edges of the associated node network of the key parts are identified, and the cascading influence network of the key parts is generated.

9. The dynamic adjustment method for a screw air compressor based on multi-sensor feedback as described in claim 7, characterized in that, The process of obtaining the air compressor adjustment parameter space includes: The critical component operation cascaded influence network is embedded into the N air compressor state prediction models for integrated prediction coupling to obtain a multi-component coupled state prediction model. Based on the output results of the N air compressor state prediction models, determine the component abnormal deviation parameter set; Based on the multi-component coupling state prediction model, the adjustment parameter range of the abnormal deviation parameter set of the components is analyzed to obtain the adjustment parameter space of the air compressor.

10. A dynamic adjustment system for a screw air compressor based on multi-sensor feedback, characterized in that, For implementing the multi-sensor feedback-based dynamic adjustment method for screw air compressors according to any one of claims 1-9, the system comprises: Platform construction module: The target screw air compressor is divided into N key parts. Based on the N key parts, a screw air compressor adjustment platform is built. The screw air compressor adjustment platform includes a multi-parameter sensor network, an edge computing unit, and an adjustment actuator. Modeling module: Collects N multi-source heterogeneous operating parameters of the N key parts of the air compressor through the multi-parameter sensor network, and performs state prediction modeling on the N multi-source heterogeneous operating parameters based on the edge computing unit to generate N air compressor state prediction models; Analysis module: Based on the multiple objectives of air compressor regulation, design the air compressor objective function, and use the air compressor objective function to perform feedback regulation analysis on the N air compressor state prediction models to determine the target air compressor regulation parameters; Dynamic adjustment module: Based on the matching of the target air compressor adjustment parameters with the adjustment actuator, the adjustment mechanism combination is activated, and the air compressor operation is dynamically adjusted according to the target air compressor adjustment parameters through the adjustment mechanism combination.

Images