Air compressor frequency conversion energy-saving control method and system based on data analysis

Abstract

The invention discloses an air compressor frequency conversion and energy saving control method and system based on data analysis, and particularly relates to the technical field of air compressor control. Aiming at the problems of frequency jitter and motor instability of the permanent magnet variable-frequency screw air compressor in a high-frequency air consumption fluctuation scene, the method comprises the following steps of: acquiring exhaust pressure, air consumption flow, motor current, frequency converter output frequency and environment temperature data in real time, performing trend prediction by utilizing a data analysis model, and constructing a multi-dimensional working condition characteristic curve; a gas consumption fluctuation mode is accurately recognized, and predictive judgment is made before frequency fluctuation is about to exceed the limit; according to a prediction result, a frequency converter control strategy is dynamically adjusted, the frequency response rate is buffered in advance, an energy-saving mode is automatically triggered when the load change trend is abnormal, rotating speed control is optimized in real time, and according to the method, the operation stability and response precision of the air compressor are effectively improved, energy consumption is remarkably reduced, and the service life of equipment is prolonged; the method is suitable for the intelligent energy-saving operation requirement in a high-beat and high-precision manufacturing scene.

Description

Technical Field

[0001] The present invention relates to the technical field of air compressor control, and particularly to a variable-frequency energy-saving control method and system for air compressors based on data analysis. Background Art

[0002] With the continuous improvement of industrial automation, air compressors, as key power equipment, are widely used in industries such as manufacturing, metallurgy, chemical engineering, and electronics. Among many types, permanent magnet variable-frequency screw air compressors have gradually become the mainstream products due to their characteristics of high efficiency, energy saving, and low noise. Compared with the traditional power-frequency drive method, variable-frequency control can dynamically adjust the speed according to the actual gas consumption, significantly reducing energy consumption and improving operating efficiency.

[0003] The prior art has the following deficiencies:

[0004] In some manufacturing workshops with concentrated high-precision pneumatic equipment, such as the SMT patch line of an electronic manufacturing enterprise, due to the rapid production rhythm and drastic changes in gas consumption, frequency jitter problems occur in the real-time response of permanent magnet variable-frequency screw air compressors, that is, the frequency of the frequency converter fluctuates violently in a short period of time, resulting in unstable operation of the motor. Long-term operation will accelerate bearing wear and cause the system to alarm and stop. Summary of the Invention

[0005] The purpose of the present invention is to provide a variable-frequency energy-saving control method and system for air compressors based on data analysis to solve the deficiencies in the background art.

[0006] To achieve the above purpose, the present invention provides the following technical solution: A variable-frequency energy-saving control method for air compressors based on data analysis, including:

[0007] Collecting real-time operation data during the operation of the air compressor, including exhaust pressure, gas flow rate, motor current, frequency converter output frequency, and ambient temperature;

[0008] Performing dynamic trend prediction on the real-time operation data through a data analysis model, constructing a multi-dimensional operating condition characteristic curve, identifying the gas consumption fluctuation pattern based on the characteristic curve, and making a predictive judgment before the frequency fluctuation amplitude exceeds the set threshold;

[0009] Adjusting the frequency converter control strategy according to the prediction result, buffering the variable-frequency response rate in advance, and triggering an adaptive energy-saving mode when detecting an abnormal load change trend, and adjusting the operating speed of the compressor in real time.

[0010] Preferably, the output pressure of compressed air is monitored in real time according to the pressure sensor set at the exhaust port of the compressor. A mass flowmeter or a vortex flowmeter is installed on the output pipeline or the main pipeline of the compressor to obtain the actual flow data of the gas-using system. The instantaneous current value during the operation of the motor is obtained through the built-in current sensing unit of the frequency converter or an externally installed Hall current sensor. The actual working frequency value output by the frequency converter to the main motor is directly read, and the working environment temperature is monitored by the temperature sensor arranged at the air inlet of the compressor or in the machine room.

[0011] Preferably, a random forest regression model is used to perform dynamic trend prediction on the multi-dimensional real-time operation data collected by the air compressor:

[0012] The original operation data collected from the air compressor system includes exhaust pressure, gas consumption flow, motor current, frequency converter output frequency, and ambient temperature; it is constructed into an input feature vector set, and samples are extracted from the historical operation data in the form of a sliding window. Each sample contains: the characteristic parameters at the current and several previous moments, and the value of the target parameter within the next time period.

[0013] Construct a training set D = {(X t , Y t+Δt )} for supervised learning, where X t is the input feature vector at time t, and Y t+Δt is the target value for the future Δt time.

[0014] Use the random forest algorithm to construct multiple decision trees, and each tree is independently trained in different training subsets and feature subsets. The training process includes:

[0015] Adopt the Bootstrap sampling method to subsample the training data;

[0016] Select the best splitting feature and splitting point at each node of each tree;

[0017] Construct a regression forest composed of multiple trees, and the final prediction result is the average of the prediction results of all trees.

[0018] Preferably, during the operation of the air compressor system, the feature vector X t at the current time point is collected in real time and input into the trained random forest model, and the predicted value Y t+Δt is output, that is, the change trend of the key parameters of the system within the next time window is predicted;

[0019] Combine multiple prediction parameters and plot them into a multi-dimensional time series curve to form a working condition trend map. The horizontal axis is time, and the vertical axis is the values of different prediction variables. A set of vector data is generated at each time point to form a working condition state trajectory.

[0020] Preferably, through the trained random forest regression model, obtain the inverter output frequency sequence F = {f1, f2,..., f m}; where f m represents the predicted frequency at the m-th time step, and m is the number of steps within the prediction time window; calculate the maximum rising slope of the continuous rising segment in the frequency sequence, and extract the steepest falling slope of the continuous falling segment in the sequence.

[0021] Preferably, normalize the maximum upward rate and the steepest falling slope so that they are both between [0, 1], and calculate the mean of the normalized maximum upward rate and the steepest falling slope, that is, obtain the comprehensive fluctuation intensity evaluation coefficient.

[0022] Preferably, compare the obtained comprehensive fluctuation intensity evaluation coefficient with a predetermined threshold. If the comprehensive fluctuation intensity evaluation coefficient is greater than the predetermined threshold, it is determined that the frequency unstable region is about to be entered. At this time, it is necessary to adjust the inverter control strategy and buffer the frequency conversion response rate in advance.

[0023] Preferably, use the comprehensive fluctuation intensity evaluation coefficient, the load change rate, and the current unit gas consumption as the input items of fuzzy logic;

[0024] Use the inverter response buffer factor, the target operating frequency adjustment factor, and the energy-saving mode switching signal as the output items of fuzzy logic;

[0025] Convert the continuous numerical values of the input variables into fuzzy linguistic variables through membership functions;

[0026] According to the established rule base, perform pattern matching on the input fuzzy quantities;

[0027] Convert the inference result into a specific control output. For the energy-saving mode switching signal, directly judge whether it is correct according to the activation degree of the fuzzy output;

[0028] Transmit the final processing result to the inverter control unit and the compressor main control system to perform frequency adjustment, response buffering, and energy-saving mode switching operations, forming a closed-loop regulation.

[0029] The present invention also provides an air compressor variable-frequency energy-saving control system based on data analysis, including a data acquisition module, a prediction analysis module, and a control adjustment module;

[0030] Data acquisition module: Collect the real-time operation data of the air compressor during operation, including exhaust pressure, gas consumption flow, motor current, inverter output frequency, and ambient temperature;

[0031] Prediction analysis module: Dynamically predict the trend of the real-time operation data through a data analysis model, construct a multi-dimensional operating condition characteristic curve, identify the gas consumption fluctuation pattern based on the characteristic curve, and make a predictive judgment before the frequency fluctuation amplitude exceeds the set threshold;

[0032] Control adjustment module: Adjust the frequency converter control strategy according to the prediction result, buffer the frequency conversion response rate in advance, and trigger the adaptive energy-saving mode when detecting an abnormal load change trend, and adjust the operating speed of the compressor in real time.

[0033] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0034] 1. By introducing a data analysis model based on random forest regression and combining multi-parameter real-time data acquisition and multi-dimensional operating condition characteristic curve construction, the present invention realizes the dynamic trend prediction and intelligent control of the operating state of the air compressor. Compared with the traditional method that relies on fixed logic or PID control, this method can identify the gas consumption fluctuation pattern in advance, make a predictive judgment before the frequency fluctuation is about to exceed the limit, adjust the frequency converter response rate in time, significantly suppress the frequency jitter, reduce the risk of motor impact and system failure, and effectively extend the service life of the equipment.

[0035] 2. The present invention further combines fuzzy logic control technology, introduces key indicators such as comprehensive fluctuation intensity, load change rate, and unit energy consumption as inputs, realizes the adaptive adjustment of the frequency conversion control strategy and the intelligent switching of the energy-saving mode, thereby dynamically optimizing the operating speed of the compressor according to the actual operating conditions, improving the energy efficiency ratio and operating stability of the system, and is especially suitable for modern manufacturing workshops with frequent gas load fluctuations and high-speed beats. The overall solution has high real-time performance, high adaptability and significant energy-saving effects, and has good engineering application prospects. Description of the Drawings

[0036] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0037] Figure 1 It is the method flow chart of the present invention.

[0038] Figure 2 It is the system module diagram of the present invention. Detailed Embodiments

[0039] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0040] Example 1. Please refer to Figure 1 As shown, a frequency conversion energy-saving control method for an air compressor based on data analysis in this embodiment includes:

[0041] Collect real-time operation data of the air compressor during operation, including exhaust pressure, gas consumption flow rate, motor current, frequency converter output frequency, and ambient temperature;

[0042] Perform dynamic trend prediction on the real-time operation data through a data analysis model, construct a multi-dimensional operating condition characteristic curve, identify the gas consumption fluctuation pattern based on the characteristic curve, and make a predictive judgment before the frequency fluctuation amplitude exceeds the set threshold;

[0043] Adjust the frequency converter control strategy according to the prediction result, buffer the frequency conversion response rate in advance, and trigger the adaptive energy-saving mode when detecting an abnormal load change trend, and adjust the operating speed of the compressor in real time.

[0044] Collect the key operating parameters of the permanent magnet variable frequency screw air compressor in real time, specifically including the following:

[0045] Set a high-precision pressure sensor (such as a pressure transmitter) at the exhaust port of the compressor to monitor the output pressure of the compressed air in real time. This parameter reflects the required air source pressure of the current system and is an important basis for adjusting the output frequency of the frequency converter. The sampling frequency can be set to 100 - 500 ms and adjusted according to the actual response requirements.

[0046] Install a mass flowmeter or a vortex flowmeter on the output pipeline or the main pipeline of the compressor to obtain the actual flow data of the gas consumption system in real time. This data is used to judge the load change of the gas consumption end and predict the short-term gas consumption trend in combination with historical data, providing a reference basis for frequency conversion control.

[0047] Obtain the instantaneous current value of the motor during operation through the built-in current sensing unit of the frequency converter or an externally installed Hall current sensor. The motor current is an important indicator for evaluating load changes and judging the operating efficiency of the system, and is also used to warn of potential problems such as overload and mechanical failures.

[0048] Directly read the actual working frequency value output by the frequency converter to the main motor, which reflects the change in the compressor speed. There is a dynamic correlation between this parameter and the exhaust pressure and gas flow rate, and it is a key adjustment variable in the entire energy-saving control model.

[0049] Install temperature sensors at the compressor air inlet or in the machine room to monitor the working environment temperature. The ambient temperature has a direct impact on the compressor efficiency and the motor heating state. The system can adjust the operating mode (such as automatic speed limit, enhanced heat dissipation, etc.) according to the temperature change to ensure the safe and efficient operation of the equipment.

[0050] All the above sensing data are connected to the data acquisition module through the PLC control system or the industrial gateway and enter the central control unit for subsequent data analysis models to perform fusion calculations, working condition identification, and predictive control.

[0051] The collected original operation data (including exhaust pressure, gas flow rate, motor current, frequency converter output frequency, ambient temperature, etc.) are first subjected to filtering, noise reduction, and standardization processing to eliminate outliers and interference signals to ensure the accuracy and stability of subsequent analysis.

[0052] The system divides data windows according to the set time intervals (such as 1 minute, 5 minutes, etc.), extracts data features within each period, such as statistical quantities like mean, maximum value, minimum value, standard deviation, etc., and calculates the correlation and co-variation trend between variables.

[0053] Introduce time series analysis (such as ARIMA model), machine learning algorithms (such as LSTM neural network, random forest regression), or industrial big data analysis methods to predict the trend of each operating parameter over time.

[0054] The present invention uses a random forest regression model to perform dynamic trend prediction on the multi-dimensional real-time operation data collected by the air compressor. The specific steps are as follows:

[0055] The original operation data collected from the air compressor system includes but is not limited to the following parameters:

[0056] Exhaust pressure (P), gas flow rate (Q), motor current (I), frequency converter output frequency (F), and ambient temperature (T); in addition, time stamps and historical trend data (such as mean, standard deviation, slope, etc. within the previous n minutes) can be introduced as auxiliary features. The above parameters are constructed into an input feature vector set X = {x1, x2,..., x n}, and the corresponding prediction target is the change trend of key parameters (such as pressure, frequency) within a future time window.

[0057] Extract samples from the historical operation data in a sliding window manner. Each sample contains:

[0058] Characteristic parameters at the current and previous moments (as inputs);

[0059] Values of target parameters within the next time period (such as the next 30 seconds or 1 minute) (as outputs);

[0060] Construct a training set D = {(X t , Y t+Δt )} for supervised learning, where X t is the input feature vector at time t, and Y t+Δt is the target value for the future Δt time (such as frequency, pressure);

[0061] Use the random forest algorithm to construct multiple decision trees, and each tree is independently trained in different training subsets and feature subsets to enhance the generalization ability. The training process includes:

[0062] Adopt the Bootstrap sampling method to subsample the training data;

[0063] Select the best splitting feature and splitting point at each node of the tree;

[0064] Construct a regression forest composed of multiple trees, and the final prediction result is the average of the prediction results of all trees.

[0065] During the operation of the air compressor system, the feature vector X t at the current time point is collected in real time and input into the trained random forest model, and the predicted value Y t+Δt is output, that is, the change trend of the key parameters of the system within a future time window is predicted.

[0066] Combine multiple predicted parameters (such as pressure, frequency, and current changes within the next 30 seconds) to plot a multi-dimensional time series curve to form a working condition trend map. The horizontal axis is time (t), and the vertical axis is the values of different predicted variables (such as P(t), F(t), I(t)). A set of vector data is generated at each time point to form a working condition state trajectory.

[0067] Continuously monitor the multiple parameters output by real-time prediction in the form of a sliding time window to generate a sequence of dynamic feature curves. For example, within each 30-second sliding window, record the change trend curves of each parameter and continuously refresh the data.

[0068] Based on historical operation data, pre-define several typical gas consumption fluctuation patterns, including:

[0069] Steady state mode: The change amplitudes of flow rate and frequency are tiny, and the compressor operates at a constant speed;

[0070] Light load periodic fluctuation: The flow rate rises and falls periodically, and the frequency jitters slightly;

[0071] Rapid load fluctuations: flow rate suddenly rises or falls, with obvious frequency response;

[0072] Abnormal fluctuation / sudden change: The frequency fluctuates sharply in a short time, which may cause potential failure risk.

[0073] The above curves are extracted using machine learning models such as K-means clustering or support vector machine (SVM), and the current characteristic curve is classified into the closest historical operating mode to achieve automatic identification.

[0074] Through the trained random forest regression model, the inverter output frequency sequence in the future prediction period is obtained: F = {f1, f2, ..., f m}; where f m Indicates the prediction frequency of the mth time step, where m is the number of steps in the prediction time window (e.g., 30 points are sampled in 30 seconds).

[0075] Calculate the maximum slope of the continuous rising segment in the frequency sequence, reflecting the intensity of the upward fluctuation of the frequency in a short period of time. It is defined as: Among them, S up Indicates the maximum uplink rate, indicating a sharp increase in frequency, t i ,t j Indicates two different sampling time points;

[0076] Similarly, the steepest slope of the consecutive descending segments in the sequence is extracted, defined as: S down It is the steepest descending slope. This value is a negative number. The larger its absolute value is, the more drastic the frequency decreases, which means the higher the risk of frequency undershoot.

[0077] The maximum upward rate and the steepest downward slope are normalized so that they are both between [0,1]. The average of the normalized maximum upward rate and the steepest downward slope is calculated to obtain the comprehensive fluctuation intensity assessment coefficient.

[0078] The obtained comprehensive fluctuation intensity assessment coefficient is compared with the preset threshold. If the comprehensive fluctuation intensity assessment coefficient is greater than the preset threshold, it is determined that the frequency is about to enter the unstable frequency zone. At this time, the inverter control strategy needs to be adjusted to buffer the frequency conversion response rate in advance.

[0079] The comprehensive fluctuation intensity assessment coefficient, load change rate and current unit gas energy consumption are used as input items of fuzzy logic;

[0080] The comprehensive fluctuation intensity assessment coefficient reflects the severity of the change in the compressor frequency in a short period of time and is an important priori signal for determining whether the system is facing frequency shocks.

[0081] The load change rate is obtained by calculating the first derivative of the gas consumption flow rate in real time, representing the change speed of the load per unit time. A positive value indicates an increasing load, a negative value indicates a decreasing load, and a zero value or a value close to zero indicates a stable system load. This parameter reflects the dynamics and fluctuation frequency of the system operating conditions.

[0082] The energy consumption per unit gas volume is obtained by dividing the current compressor power by the output gas volume, reflecting the current energy efficiency level of the compressor. When this value is high, it indicates that the system operation deviates from the optimal efficiency range, which is an important basis for starting energy-saving strategies.

[0083] The inverter response buffer factor, the target operating frequency adjustment factor, and the energy-saving mode switching signal are used as the output items of the fuzzy logic.

[0084] The inverter response buffer factor (α) is used to adjust the acceleration and deceleration slope of the inverter, that is, the speed at which the inverter responds to changes. When α is high, it indicates that the frequency response needs to be buffered, and the controller automatically reduces the frequency change rate of the inverter to avoid mechanical shocks or system instability caused by too fast response.

[0085] The target operating frequency adjustment factor (β) is used to correct the target operating frequency of the compressor, so as to dynamically fit the actual gas consumption demand and the optimal energy efficiency point. When the prediction shows that the system load will decrease, β outputs a negative adjustment, and the compressor frequency can be appropriately reduced; conversely, if the prediction shows that the load is about to increase, β is a positive adjustment to increase the frequency in advance.

[0086] The energy-saving mode switching signal is a Boolean logic output, used to judge whether to immediately switch to the energy-saving operation mode. When the system load changes slowly, the energy consumption is high, or the fluctuation is severe but there is no need for rapid response, the system automatically enters the energy-saving mode, such as reducing the operating frequency, starting intermittent compression, and extending the shutdown time.

[0087] The fuzzy controller realizes the decision-making process through the following reasoning steps:

[0088] The continuous numerical values of the input variables are converted into fuzzy linguistic variables through membership functions. For example, the comprehensive fluctuation intensity can be fuzzified into "low", "medium", "high", "extremely high"; the load change rate is fuzzified into "rapid decline", "slow change", "rapid increase"; the energy consumption per unit gas volume is fuzzified into "high consumption", "medium consumption", "low consumption".

[0089] The system performs pattern matching on the input fuzzy quantities according to the established rule base. For example, when "the fluctuation intensity is high", "the load is increasing rapidly", and "the unit energy consumption is medium consumption", the controller will judge that the system is about to face a dynamic load impact and has not reached the efficient operation state, and should moderately buffer the response, increase the target frequency at the same time, and not trigger the energy-saving mode temporarily.

[0090] For another example, when the "fluctuation intensity is medium", the "load change is gentle", and the "energy consumption is high", the controller will infer that the current system is operating redundantly, and can slow down the frequency conversion response speed, lower the operating frequency, and immediately switch to the energy-saving mode to reduce energy consumption.

[0091] Convert the inference result into specific control outputs. For α and β, the centroid method is used for numerical defuzzification, and the outputs are specific buffer factors and frequency adjustment amounts (for example, α = 0.7 means that the output change speed limit is 70%; β = -2 means that the target frequency is lowered by 2 Hz). For the energy-saving mode switching signal, it is directly judged whether it is correct according to the activation degree of the fuzzy output.

[0092] Transmit the final processing result to the frequency converter control unit and the compressor main control system, and perform operations such as frequency adjustment, response buffering, and energy-saving mode switching to form a closed-loop regulation.

[0093] Example 2, please refer to Figure 2 As shown, the air compressor variable-frequency energy-saving control system based on data analysis in this embodiment includes a data acquisition module, a prediction analysis module, and a control adjustment module;

[0094] Data acquisition module: Collect real-time operation data of the air compressor during operation, including exhaust pressure, gas consumption flow, motor current, frequency converter output frequency, and ambient temperature;

[0095] Prediction analysis module: Dynamically predict the real-time operation data through a data analysis model, construct a multi-dimensional working condition characteristic curve, identify the gas consumption fluctuation pattern based on the characteristic curve, and make a predictive judgment before the frequency fluctuation amplitude exceeds the set threshold;

[0096] Control adjustment module: Adjust the frequency converter control strategy according to the prediction result, buffer the frequency conversion response rate in advance, and trigger the adaptive energy-saving mode when detecting an abnormal load change trend, and adjust the compressor operating speed in real time.

[0097] The above formulas are all dimensionless and take their numerical calculations. The formula is a formula obtained by software simulation of collecting a large amount of data to get the closest to the real situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0098] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. These three situations, where A and B can be singular or plural. In addition, the character " / " in this article generally represents an "or" relationship between the front and back associated objects, but it may also represent an "and / or" relationship, which can be specifically understood with reference to the context.

[0099] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware or in combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0100] As described above, the above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed by this application can easily think of changes or substitutions, and all should be covered within the protection scope of this application.

Claims

1. A variable-frequency energy-saving control method for an air compressor based on data analysis, characterized in that: Including: Collecting the real-time operation data of the air compressor during operation, including exhaust pressure, gas consumption flow rate, motor current, frequency converter output frequency, and ambient temperature; Performing dynamic trend prediction on the real-time operation data through a data analysis model, constructing a multi-dimensional operating condition characteristic curve, identifying the gas consumption fluctuation pattern based on the characteristic curve, and making a predictive judgment before the frequency fluctuation amplitude exceeds the set threshold; Adjusting the frequency converter control strategy according to the prediction result, buffering the frequency conversion response rate in advance, and triggering the adaptive energy-saving mode when detecting an abnormal load change trend, and adjusting the operating speed of the compressor in real time.

2. The variable frequency energy-saving control method for an air compressor based on data analysis according to claim 1, characterized in that: Real-time monitoring of the output pressure of compressed air based on the pressure sensor set at the exhaust port of the compressor, installing a mass flowmeter or vortex flowmeter on the output pipeline or main pipeline of the compressor to obtain the actual flow data of the gas consumption system, obtaining the instantaneous current value during motor operation through the built-in current sensing unit of the frequency converter or an externally installed Hall current sensor, directly reading the actual working frequency value output by the frequency converter to the main motor, and monitoring the working environment temperature through the temperature sensor arranged at the air inlet of the compressor or in the machine room.

3. The air compressor variable frequency energy-saving control method based on data analysis according to claim 2, characterized in that: Using a random forest regression model to perform dynamic trend prediction on the multi-dimensional real-time operation data collected by the air compressor: The original operation data collected from the air compressor system, including exhaust pressure, gas consumption flow rate, motor current, frequency converter output frequency, and ambient temperature; constructing it into an input feature vector set, extracting samples from the historical operation data in a sliding window manner, and each sample contains: the characteristic parameters at the current and previous several moments, and the value of the target parameter in the next time period; Construct a training set D = {(X t , Y t+Δt )} for supervised learning, where X t is the input feature vector at time t, and Y t+Δt is the target value for the future Δt time; Using the random forest algorithm to construct multiple decision trees, and each tree is independently trained in different training subsets and feature subsets. The training process includes: Sub-sampling the training data using the Bootstrap sampling method; Selecting the best splitting feature and splitting point at each node of the tree; Constructing a regression forest composed of multiple trees, and the final prediction result is the average of the prediction results of all trees.

4. A frequency conversion energy-saving control method for an air compressor based on data analysis according to claim 3, characterized in that: During the operation of the air compressor system, the feature vector X at the current time point is collected in real time t , and input into the trained random forest model to output the predicted value Y t+Δt , that is, to predict the change trend of the key parameters of the system within a future time window; Combining multiple prediction parameters and plotting them into a multi-dimensional time series curve to form an operating condition trend map. The horizontal axis is time, the vertical axis is the values of different prediction variables, and a set of vector data is generated at each time point to form an operating condition state trajectory.

5. A variable frequency energy-saving control method for an air compressor based on data analysis according to claim 4, characterized in that: Through the trained random forest regression model, obtain the inverter output frequency sequence F = {f1, f2,..., f m} within the future prediction period; where f m represents the predicted frequency at the m-th time step, and m is the number of steps within the prediction time window; calculate the maximum rising slope of the continuous rising segments in the frequency sequence, and extract the steepest falling slope of the continuous falling segments in the sequence.

6. A frequency conversion energy-saving control method for an air compressor based on data analysis according to claim 5, characterized in that: Normalizing the maximum upward rate and the steepest downward slope so that they are both between [0, 1], and calculating the mean value of the normalized maximum upward rate and the steepest downward slope to obtain the comprehensive fluctuation intensity evaluation coefficient.

7. A frequency conversion energy-saving control method for an air compressor based on data analysis according to claim 6, characterized in that: Comparing the obtained comprehensive fluctuation intensity evaluation coefficient with a predetermined threshold. If the comprehensive fluctuation intensity evaluation coefficient is greater than the predetermined threshold, it is determined that it is about to enter the frequency unstable area. At this time, it is necessary to adjust the frequency converter control strategy to buffer the frequency conversion response rate in advance.

8. A frequency conversion energy-saving control method for an air compressor based on data analysis according to claim 7, characterized in that: Taking the comprehensive fluctuation intensity evaluation coefficient, the load change rate, and the current unit gas consumption energy consumption as the input items of fuzzy logic; Taking the frequency converter response buffer factor, the target operating frequency adjustment factor, and the energy-saving mode switching signal as the output items of fuzzy logic; Converting the continuous numerical values of the input variables into fuzzy linguistic variables through the membership function; Performing pattern matching on the input fuzzy quantities according to the established rule base; Convert the inference result into a specific control output. For the energy-saving mode switching signal, directly judge whether it is correct according to the activation degree of the fuzzy output; Transmit the final processing result to the frequency converter control unit and the compressor main control system, perform frequency adjustment, response buffering and energy-saving mode switching operations, and form a closed-loop regulation.

9. A variable-frequency energy-saving control system for an air compressor based on data analysis, which is used to implement a variable-frequency energy-saving control method for an air compressor based on data analysis according to any one of claims 1-8, characterized in that: It includes a data acquisition module, a prediction analysis module and a control adjustment module; Data acquisition module: Collect the real-time operation data of the air compressor during operation, including exhaust pressure, gas consumption flow, motor current, frequency converter output frequency and ambient temperature; Prediction analysis module: Dynamically predict the trend of the real-time operation data through a data analysis model, construct a multi-dimensional working condition characteristic curve, identify the gas consumption fluctuation mode based on the characteristic curve, and make a predictive judgment before the frequency fluctuation amplitude exceeds the set threshold; Control adjustment module: Adjust the frequency converter control strategy according to the prediction result, buffer the frequency conversion response rate in advance, and trigger the adaptive energy-saving mode when detecting an abnormal load change trend, and adjust the operating speed of the compressor in real time.

Images