A switching control method for operation mode of air compressor

Abstract

This invention discloses a switching control method for air compressor operating modes, specifically relating to the field of industrial air compressor unit control technology. The method involves acquiring pressure signals and analyzing their frequency and time domain characteristics to output fluctuation level parameters, thereby dynamically adjusting the pressure switching dead zone threshold and delay time. It also involves detecting the frequency of the variable frequency drive (VFD), calculating the compensation amount using a piecewise compensation function, and superimposing it onto the PID output to correct the pressure. Furthermore, it monitors slave unit unloading and pressure rebound, generating a virtual locking signal to prohibit all second slave units from unloading within a specified window. The method continuously calculates the absolute and net integrals of the error value, and when the absolute integral exceeds the threshold, it forcibly triggers a re-optimization of the scheduling strategy. This invention avoids frequent start-stop of slave units by dynamically adjusting the dead zone through fluctuation identification, corrects nonlinear distortion of the VFD through piecewise compensation, and locks the second slave unit to cut off the interlocking unloading during pressure rebound. Combined with error integral re-optimization parameters, this improves the air supply stability and equipment lifespan under high-frequency fluctuation conditions.

Description

Technical Field

[0001] This invention relates to the field of industrial air compressor unit control technology, and more specifically, to a method for switching operating modes of an air compressor. Background Technology

[0002] Air compressor stations widely employ a configuration where a variable frequency drive (VFD) unit and multiple line-frequency slave units operate in parallel, regulating the air supply pressure through a master-slave linkage strategy. When the air load is stable, conventional proportional-integral-derivative (PI-DE) control with a fixed dead zone can maintain stable pipeline pressure. However, in actual industrial production, such as automotive welding, intermittent pneumatic tool operation, and semiconductor cleanroom purging, the air load often exhibits high-frequency intermittent fluctuations, meaning that the pressure repeatedly jumps significantly within seconds to tens of seconds. Under these conditions, the VFD unit exhibits significant nonlinearity in the flow-pressure relationship at extremely low speeds or near full load, causing conventional PI-DE control to lag or overshoot.

[0003] Frequent start-stop cycles of the power frequency slave units in a high-frequency fluctuation environment not only accelerate the mechanical wear of components such as contactors and intake valves, but also easily trigger a momentary pressure rebound due to the pipeline cavity effect at the instant of unloading of one slave unit, thereby inducing multiple slave units to unload in a chain reaction. This chain reaction can cause a precipitous drop in gas supply capacity, and may even lead to a complete loss of pressure and production interruption, making it difficult to guarantee the stability of the switching process and the reliability of gas supply.

[0004] Furthermore, after long-term operation, air compressor units will experience accumulated pressure deviations due to wear or changes in operating conditions. Existing control strategies mostly rely on fixed parameters or simple periodic corrections, which limits the adaptability of the air compressor unit under high-frequency fluctuations. This often requires operators to manually adjust parameters frequently, impacting the overall economy and automation of the air compressor station. Therefore, this invention proposes a switching control method for air compressor operating modes to address the aforementioned problems. Summary of the Invention

[0005] To achieve the above objectives, the present invention provides the following technical solution: A method for switching operating modes of an air compressor, comprising the following steps: The pressure signal in the gas supply pipeline is collected in real time. Frequency domain and time domain analysis is performed on the pressure signal to calculate the rate of change of pressure fluctuation amplitude and the frequency components of fluctuation. Based on the rate of change of pressure fluctuation amplitude and the frequency components of fluctuation, the fluctuation state of the current gas load is determined and the fluctuation level parameter is output. The preset pressure switching dead zone threshold in the linkage control of multiple air compressors is dynamically adjusted according to the fluctuation level parameter, and the delay confirmation time is also adjusted. The operating frequency of the current inverter host is obtained. When the operating frequency is less than the first frequency threshold or greater than the second frequency threshold, the compensation amount is calculated according to the piecewise compensation function and the compensation amount is superimposed on the output value of the proportional-integral-derivative controller to correct the deviation between the gas supply pressure and the target pressure. The first frequency threshold is less than the second frequency threshold. Monitor the loading and unloading action event sequence of all air compressor units and the instantaneous pressure rebound characteristics after unloading. When the air supply pressure rebounds after the first slave unit is unloaded and the second slave unit meets the condition of the unloading threshold based on the preset pressure switching dead zone threshold, generate a virtual lock signal and temporarily prohibit all determined second slave units from performing unloading actions within the preset time window. The error between the actual gas supply pressure and the target pressure is continuously calculated, and the absolute value of the error is integrated over time to obtain the cumulative error. At the same time, the error value itself is integrated over time to obtain the net cumulative error. When the cumulative error exceeds the preset error threshold, the re-optimization of the scheduling strategy is forcibly triggered.

[0006] The fluctuation level parameters include a first level representing non-high-frequency intermittent fluctuation states and a second level representing high-frequency intermittent fluctuation states.

[0007] In a preferred embodiment, the specific process of determining the fluctuation state of the current gas load based on the rate of change of pressure fluctuation amplitude and the frequency components of the fluctuation includes: Pressure time series is obtained by continuously collecting pressure signals in the gas supply pipeline at preset sampling intervals. Pressure data within a sliding window before the current moment is taken. The length of the sliding window is a preset number of sampling points. Calculate the difference between the maximum and minimum pressure values ​​within the sliding window, divide it by the total duration of the sliding window, and obtain the rate of change of pressure fluctuation amplitude. Perform a fast Fourier transform on the pressure signal within the sliding window and extract the frequency component with the largest amplitude as the main frequency. Compare the main frequency with the preset frequency threshold value, and compare the pressure fluctuation amplitude change rate with the preset amplitude threshold value. When the main frequency is greater than the frequency threshold and the rate of change of pressure fluctuation amplitude is greater than the amplitude threshold, the output is the second level fluctuation level parameter representing the high-frequency intermittent fluctuation state; otherwise, the output is the first level fluctuation level parameter representing the non-high-frequency intermittent fluctuation state.

[0008] The pressure switching dead zone threshold includes a first pressure threshold for starting the slave and a second pressure threshold for unloading the slave.

[0009] In a preferred embodiment, the specific process of dynamically adjusting the pressure switching dead zone threshold in the multi-air compressor linkage control based on the fluctuation level parameter, while simultaneously adjusting the delay confirmation time, includes: The baseline values ​​for the first pressure threshold, the second pressure threshold, and the delayed confirmation time are preset. When the fluctuation level parameter is the first level representing a non-high-frequency intermittent fluctuation state, the first pressure threshold and the second pressure threshold are kept as their respective baseline values, and the delayed confirmation time is set as the baseline value. When the fluctuation level parameter is the second level representing a high-frequency intermittent fluctuation state, the first pressure threshold is reduced from the reference value by a first preset ratio, the second pressure threshold is increased from the reference value by a second preset ratio, and the delay confirmation time is increased from the reference value by a preset multiple.

[0010] In a preferred embodiment, the specific process of calculating the compensation amount based on the piecewise compensation function includes: The interval from zero frequency to the first frequency threshold is pre-divided into a first sub-interval, a second sub-interval, and a third sub-interval, and the interval from the second frequency threshold to the maximum operating frequency is pre-divided into a fourth sub-interval and a fifth sub-interval. Obtain the current operating frequency of the inverter host. When the operating frequency is greater than or equal to the first frequency threshold and less than or equal to the second frequency threshold, the compensation amount is zero. When the operating frequency is less than the first frequency threshold, the compensation amount is calculated according to the frequency sub-intervals into which the operating frequency falls: When the frequency falls into the first sub-interval, the compensation amount is equal to the first proportional coefficient multiplied by the square of the difference between the operating frequency and the first frequency threshold. When the value falls into the second subinterval, the compensation amount is equal to the second proportional coefficient multiplied by the cube of the difference; when the value falls into the third subinterval, the compensation amount is equal to the third proportional coefficient multiplied by the fourth power of the difference. When the operating frequency is greater than the second frequency threshold, the compensation amount is calculated according to the frequency sub-intervals into which the operating frequency falls: When the frequency falls into the fourth sub-interval, the compensation amount is equal to the fourth proportional coefficient multiplied by the sine value of the difference between the operating frequency and the second frequency threshold. When the value falls into the fifth subinterval, the compensation amount is equal to the fifth proportional coefficient multiplied by the exponential function value of the difference. The exponential function value is equal to the difference of the natural constant raised to the power of one. When the compensation is positive, it is added to the output value of the proportional-integral-derivative controller; when the compensation is negative, its absolute value is subtracted from the output value.

[0011] In a preferred embodiment, the first scaling factor, the second scaling factor, the third scaling factor, the fourth scaling factor, and the fifth scaling factor are pre-calibrated according to the following process: Within the range of no-load to full-load of the inverter host, select multiple test frequency points and measure the actual output flow at each test frequency point; For each test frequency point, the theoretical linear flow rate corresponding to that frequency point is calculated based on the direct proportional relationship between frequency and flow rate. Specifically, the test frequency point is divided by the rated frequency of the inverter host and then multiplied by the rated flow rate. Divide the difference between the actual output flow rate and the theoretical linear flow rate at each test frequency point by the theoretical linear flow rate to obtain the nonlinear distortion rate at that frequency point. The curves of distortion rate versus frequency within the range from zero frequency to the first frequency threshold are fitted to obtain the first proportional coefficient, the second proportional coefficient, and the third proportional coefficient, respectively. The first proportional coefficient corresponds to the average distortion rate within the first frequency sub-interval, the second proportional coefficient corresponds to the average distortion rate within the second frequency sub-interval, and the third proportional coefficient corresponds to the average distortion rate within the third frequency sub-interval. The distortion rate curves within the range from the second frequency threshold to the maximum operating frequency are fitted to obtain the fourth and fifth proportional coefficients, respectively. The fourth proportional coefficient corresponds to the average distortion rate within the fourth frequency sub-interval, and the fifth proportional coefficient corresponds to the average distortion rate within the fifth frequency sub-interval.

[0012] In a preferred embodiment, generating a virtual lock signal refers to: A pre-established loading / unloading action event queue is used to record the time of unloading action of each slave unit and the gas supply pressure value within one sampling cycle before and after unloading in real time. When any slave device completes the unloading action, the gas supply pressure value after unloading is continuously collected in the event queue. The maximum pressure increment within the first preset time after unloading is calculated. If the maximum pressure increment exceeds the preset rebound threshold, it is determined that an instantaneous pressure rebound feature has occurred, and the slave device is marked as the first slave device. Iterate through all slave devices that meet the unloading condition based on the second pressure threshold, and exclude the first slave device to obtain the set of candidate slave devices; A prohibited quantity is dynamically determined based on the maximum pressure increment and the deviation between the current gas supply pressure and the target pressure. Select a prohibited number of slave devices from the set of candidate slave devices according to their priority, and mark each selected slave device as the second slave device. A virtual locking signal is generated for each second slave device. The virtual locking signal contains the identity of the corresponding second slave device and a locking time window. The duration of the locking time window is a preset second preset duration. During the locked time window, all second slave devices are prohibited from responding to any unload commands based on the second pressure threshold; After the lockout time window ends, all virtual lockout signals will be automatically cleared, and the normal unloading function of all second slave devices will be restored.

[0013] In a preferred embodiment, the specific process for determining the prohibited quantity is as follows: Pre-set a reference bounce step size, a reference deviation step size, and a scaling factor; The first ratio of the maximum pressure increment to the baseline rebound step size is calculated, and the second ratio of the absolute value of the deviation between the current gas supply pressure and the target pressure to the baseline deviation step size is calculated. The larger of the first and second ratios is taken, and the larger value is multiplied by the scaling factor to obtain the initial prohibition quantity. The initial number of prohibited items is rounded up to obtain the rounded number of prohibited items. The number of prohibited items after rounding is limited to the range of the preset maximum number of prohibited items, and is used as the number of prohibited items.

[0014] In a preferred embodiment, forcibly triggering the re-optimization of the scheduling strategy refers to: The re-optimization includes correcting the target pressure setpoint of the proportional-integral-derivative controller, reallocating the master-slave ratio, and adjusting the baseline values ​​of the first and second pressure thresholds, specifically: Increase the target pressure by the net cumulative error multiplied by the first preset value; Calculate the ratio of the arithmetic mean of the running time of each slave to the fourth root mean difference, then multiply it by the ratio of the median to the maximum absolute deviation, and finally multiply it by a preset conversion value to obtain the correction value, where the maximum absolute deviation is the maximum value of the absolute value of the difference between the running time of each slave and the median. Multiply the second preset value by the correction value and then by the cumulative error to obtain an adjustment pressure value; The adjustment pressure value is used to lower the reference value of the first pressure threshold, and the same adjustment pressure value is used to raise the reference value of the second pressure threshold.

[0015] The technical effects and advantages of this invention are as follows: This invention accurately outputs fluctuation level parameters by real-time acquisition of pressure signals and performing frequency and time domain analysis. Based on these parameters, it dynamically adjusts the pressure switching dead zone threshold and delay confirmation time. This effectively distinguishes between stable operating conditions and high-frequency intermittent fluctuation conditions, proactively expanding the dead zone range when fluctuations occur. This prevents the slave unit from frequently loading and unloading due to instantaneous pressure jumps, thereby significantly reducing the number of operations of mechanical components such as contactors and intake valves, extending the service life of the air compressor unit, and reducing the additional energy consumption caused by frequent start-stop cycles.

[0016] This invention, when the frequency converter operates in extreme regions outside of a first or second frequency threshold, calculates a compensation amount based on a piecewise compensation function and superimposes it onto the output of the proportional-integral-derivative controller, achieving precise correction of flow-pressure nonlinear distortion in low-speed and high-speed regions. This solves the problem of lag or overshoot in the response of traditional linear control in such regions, enabling the gas supply pressure to quickly and smoothly follow the target pressure, significantly improving the regulation quality of the frequency converter across the entire frequency range, and is particularly suitable for industrial sites with drastic fluctuations in gas load.

[0017] This invention monitors the loading and unloading sequence of all air compressor units and the instantaneous pressure rebound characteristics after unloading. When a pressure rebound is detected after the first slave unit unloads and a second slave unit that meets the unloading conditions exists, a virtual locking signal is generated, and all second slave units are temporarily prohibited from unloading within a preset time window, thereby breaking the causal chain of cascading unloading. By combining the absolute value integration and net integration of the error values ​​between the actual supply pressure and the target pressure, when the accumulated error exceeds a threshold, a forced re-optimization is triggered, automatically correcting the baseline values ​​of the target pressure setpoint, master-slave ratio, and pressure switching dead zone threshold. This allows the system to adaptively eliminate accumulated deviations after long-term operation, avoiding frequent manual intervention and improving the stability and automation level of the air compressor station group control system. Attached Figure Description

[0018] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings; Figure 1 This is a schematic diagram of a method for switching operating modes of an air compressor according to the present invention. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0020] Reference Figure 1 The following examples were obtained: Example 1: A method for switching operating modes of an air compressor, comprising the following steps: The pressure signal in the gas supply pipeline is collected in real time. Frequency domain and time domain analysis is performed on the pressure signal to calculate the rate of change of pressure fluctuation amplitude and the frequency components of fluctuation. Based on the rate of change of pressure fluctuation amplitude and the frequency components of fluctuation, the fluctuation state of the current gas load is determined and the fluctuation level parameter is output. The preset pressure switching dead zone threshold in the linkage control of multiple air compressors is dynamically adjusted according to the fluctuation level parameter, and the delay confirmation time is also adjusted. The operating frequency of the current inverter host is obtained. When the operating frequency is less than the first frequency threshold or greater than the second frequency threshold, the compensation amount is calculated according to the piecewise compensation function and the compensation amount is superimposed on the output value of the proportional-integral-derivative controller to correct the deviation between the gas supply pressure and the target pressure. The first frequency threshold is less than the second frequency threshold. Monitor the loading and unloading action event sequence of all air compressor units and the instantaneous pressure rebound characteristics after unloading. When the air supply pressure rebounds after the first slave unit is unloaded and the second slave unit meets the condition of the unloading threshold based on the preset pressure switching dead zone threshold, generate a virtual lock signal and temporarily prohibit all determined second slave units from performing unloading actions within the preset time window. The error between the actual gas supply pressure and the target pressure is continuously calculated, and the absolute value of the error is integrated over time to obtain the cumulative error. At the same time, the error value itself is integrated over time to obtain the net cumulative error. When the cumulative error exceeds the preset error threshold, the re-optimization of the scheduling strategy is forcibly triggered.

[0021] The fluctuation level parameters include a first level representing non-high-frequency intermittent fluctuation states and a second level representing high-frequency intermittent fluctuation states.

[0022] The specific process for determining the fluctuation state of the current gas load based on the rate of change of pressure fluctuation amplitude and the frequency components of the fluctuation includes: Pressure time series is obtained by continuously collecting pressure signals from the air supply pipeline at preset sampling intervals. Pressure data within a sliding window preceding the current moment is then taken, with the length of the sliding window being a preset number of sampling points. For example, if the sampling interval is set to 0.1 seconds and the sliding window length is set to 128 sampling points, the total window duration is 12.8 seconds, which can cover the common fluctuation cycle of air compressors (generally 5 to 15 seconds).

[0023] The difference between the maximum and minimum pressure values ​​within the sliding window is calculated and divided by the total duration of the sliding window to obtain the rate of change of pressure fluctuation amplitude. This value represents the drastic degree of pressure change per unit time. For example, if the pressure rises from 0.65 MPa to 0.75 MPa within the window, with a difference of 0.1 MPa and a total duration of 12.8 seconds, then the rate of change of amplitude is 0.0078 MPa per second. An amplitude threshold of 0.008 MPa per second is set because numerous experiments have shown that when the rate of change of amplitude exceeds 0.008, the gas load has entered a significant high-frequency intermittent fluctuation range.

[0024] A Fast Fourier Transform (FFT) is performed on the pressure signal within the sliding window, and the frequency component with the largest amplitude is extracted as the dominant frequency. The FFT converts the pressure signal from the time domain to the frequency domain, accurately identifying the dominant frequency of the fluctuation. For example, if the pressure within the window completes a full rise or fall within 6.4 seconds, then the dominant frequency is 0.156 Hz. A frequency threshold of 0.15 Hz is set because the mechanical action response time of most air compressor group control systems is approximately 6 to 7 seconds; fluctuations with frequencies higher than 0.15 Hz (i.e., periods less than 6.7 seconds) will exceed the adjustment capability of conventional controllers.

[0025] The main frequency is compared with a preset frequency threshold, and the rate of change of pressure fluctuation amplitude is compared with a preset amplitude threshold. Both of these thresholds are empirical values ​​obtained through multiple field tests. Generally, the frequency threshold is a number between 0.12 and 0.18 Hz, and the amplitude threshold is a number between 0.006 and 0.010 MPa per second.

[0026] When the dominant frequency is greater than a frequency threshold and the rate of change of pressure fluctuation amplitude is greater than an amplitude threshold, the system outputs a second-level fluctuation level parameter indicating a high-frequency intermittent fluctuation state; otherwise, it outputs a first-level fluctuation level parameter indicating a non-high-frequency intermittent fluctuation state. For example, if the dominant frequency is 0.2 Hz and the rate of change of amplitude is 0.009 MPa per second, both exceeding the threshold, the system determines that it is currently in a high-frequency intermittent fluctuation state and outputs a second-level parameter. Conversely, if the dominant frequency is 0.1 Hz or the rate of change of amplitude is 0.005 MPa per second, it outputs a first-level parameter. This allows for accurate differentiation between stable operating conditions and high-frequency fluctuation conditions requiring special control.

[0027] The pressure switching deadband threshold includes a first pressure threshold for starting the slave unit and a second pressure threshold for unloading the slave unit. The specific process of dynamically adjusting the pressure switching deadband threshold in the multi-air compressor linkage control based on fluctuation level parameters, while simultaneously adjusting the delay acknowledgment time, includes: The pressure switching dead zone threshold includes a first pressure threshold for starting a slave device and a second pressure threshold for unloading a slave device. The first pressure threshold is the minimum pressure value required for a stopped slave device to begin loading, and the second pressure threshold is the maximum pressure value required for a running slave device to stop loading. For example, if the target pressure is 0.70 MPa, the first pressure threshold is typically set to 0.68 MPa and the second pressure threshold to 0.73 MPa. This way, when the pressure drops to 0.68 MPa, a standby slave device will be started, and when the pressure rises to 0.73 MPa, a working slave device will be unloaded.

[0028] The specific process of dynamically adjusting the pressure switching dead zone threshold in the multi-air compressor linkage control based on the fluctuation level parameter, and simultaneously adjusting the delay confirmation time, includes the following operations: Pre-set the reference values ​​for the first pressure threshold, the second pressure threshold, and the delay confirmation time. For example, under stable air usage conditions, the reference value for the first pressure threshold is set to 0.68 MPa, the reference value for the second pressure threshold is set to 0.73 MPa, and the reference value for the delay confirmation time is set to 1 second. The reason for this setting is that most pneumatic equipment can operate normally in the pressure range of 0.68 to 0.73 MPa, and a 1-second delay is sufficient to filter out occasional pressure spikes.

[0029] When the fluctuation level parameter is set to the first level, representing a non-high-frequency intermittent fluctuation state, the first and second pressure thresholds are kept at their respective baseline values, and the delay confirmation time is set to the baseline value. In other words, under stable or slowly changing loads, no control parameters are changed, allowing the air compressor to operate within its normal dead zone range. This ensures response speed and avoids unnecessary switching due to unnecessary adjustments.

[0030] When the fluctuation level parameter is the second level, representing a high-frequency intermittent fluctuation state, the first pressure threshold is decreased by a first preset percentage from the reference value, the second pressure threshold is increased by a second preset percentage from the reference value, and the delay confirmation time is increased by a preset multiple from the reference value. The purpose of this operation is to expand the dead zone range and extend the judgment time, avoiding frequent start-up and shutdown of the slave device during high-frequency fluctuations. For example, if the first preset percentage is 5%, then the first pressure threshold decreases from 0.68 MPa to 0.646 MPa (0.68 multiplied by 0.95). If the second preset percentage is 5%, the second pressure threshold increases from 0.73 MPa to 0.7665 MPa (0.73 multiplied by 1.05), and if the preset multiple is 0.5 times, the delay confirmation time increases from 1 second to 1.5 seconds.

[0031] The first preset ratio, the second preset ratio, and the preset multiple are all determined based on on-site commissioning experience, generally ranging from 2% to 10%, and the multiple is between 0.3 and 0.8 times. These values ​​must be selected to ensure that the expanded dead zone does not exceed the minimum working pressure allowed by the process (e.g., 0.60 MPa) and the opening pressure of the pipeline safety valve (e.g., 0.85 MPa). Increasing the delay too much will reduce system sensitivity, while increasing it too little will not effectively filter false jumps. Therefore, a preset multiple of 0.5 times is usually a compromise value.

[0032] The specific process of calculating the compensation amount based on the piecewise compensation function includes: The specific process of calculating the compensation amount based on the piecewise compensation function includes the following operations: First, the interval from zero frequency to the first frequency threshold is divided into three equal sub-intervals: the first, second, and third sub-intervals. Then, the interval from the second frequency threshold to the maximum operating frequency is divided into two equal sub-intervals: the fourth and fifth sub-intervals. This division is because the nonlinear characteristics of the inverter are different in the low-speed and high-speed ranges, and the distortion degree varies in different depth intervals. Equal division is simple and ensures relatively consistent distortion characteristics within each sub-interval. For example, if the first frequency threshold is 25 Hz, then 0 to 25 Hz is divided into three equal sub-intervals: 0 to 8.33 Hz, 8.33 to 16.67 Hz, and 16.67 to 25 Hz. If the second frequency threshold is 45 Hz and the maximum operating frequency is 50 Hz, then 45 to 50 Hz is divided into two equal sub-intervals: 45 to 47.5 Hz and 47.5 to 50 Hz.

[0033] The current operating frequency of the inverter is obtained. When the operating frequency is greater than or equal to the first frequency threshold and less than or equal to the second frequency threshold, the compensation amount is zero. This is because within the range between the first and second frequency thresholds, the flow rate of the inverter is close to linear with the frequency, and the proportional-integral-derivative controller can already regulate the pressure well without additional compensation. For example, if the operating frequency is 35 Hz, which falls within the range of 25 to 45 Hz, the compensation amount is directly set to zero.

[0034] When the operating frequency is less than the first frequency threshold, the compensation amount is calculated based on the frequency sub-interval the operating frequency falls into. If it falls into the first sub-interval, the compensation amount is equal to the first proportional coefficient multiplied by the square of the difference between the operating frequency and the first frequency threshold. If it falls into the second sub-interval, the compensation amount is equal to the second proportional coefficient multiplied by the cube of the difference. If it falls into the third sub-interval, the compensation amount is equal to the third proportional coefficient multiplied by the fourth power of the difference. The square, cube, and fourth power functions are chosen because in the low-speed region, as the frequency decreases, the flow-pressure distortion increases at an accelerated rate. Using functions with progressively increasing powers better matches the actual distortion curve. For example, if the first frequency threshold is 25 Hz and the operating frequency is 10 Hz, it falls into the first sub-interval (outside of 0 to 8.33 Hz? Note: 10 Hz belongs to the second sub-interval of 8.33-16.67 Hz; a correct example is required). Example: Suppose the first sub-interval is 0 to 8.33 Hz, the operating frequency is 5 Hz, the difference is 20 Hz, squared and multiplied by the first proportional coefficient. The second sub-interval is from 8.33 to 16.67 Hz, with an operating frequency of 12 Hz and a difference of 13 Hz, which is then cubed and multiplied by the second scaling factor. The third sub-interval is from 16.67 to 25 Hz, with an operating frequency of 20 Hz and a difference of 5 Hz, which is then raised to the fourth power and multiplied by the third scaling factor. These power functions can compensate for the different nonlinear intensities in the shallow, middle, and deep low-velocity regions, respectively.

[0035] When the operating frequency exceeds the second frequency threshold, the compensation amount is calculated based on the frequency sub-intervals into which the operating frequency falls. When it falls into the fourth sub-interval, the compensation amount equals the fourth proportional coefficient multiplied by the sine of the difference between the operating frequency and the second frequency threshold. When it falls into the fifth sub-interval, the compensation amount equals the fifth proportional coefficient multiplied by the exponential function value of the difference. The exponential function value is equal to the difference raised to the power of the natural constant minus one. The sine function is chosen because in the high-frequency region near the second frequency threshold, distortion exhibits periodic fluctuations (similar to resonance), and the sine value can simulate this fluctuation characteristic. The exponential function is chosen because in the extremely high region near the maximum operating frequency, distortion increases sharply, and the exponential function can quickly amplify the compensation amount, preventing overshoot. For example: the second frequency threshold is 45 Hz, the fourth sub-interval is 45 to 47.5 Hz, and the operating frequency is 46.5 Hz, with a difference of 1.5 Hz. The sine of this difference is multiplied by the fourth proportional coefficient. The fifth sub-interval is from 47.5 to 50 Hz, with an operating frequency of 49 Hz and a difference of 4 Hz. The exponential function value is (e to the power of 4 minus 1), which is approximately 53.6. Multiplying this by the fifth scaling factor allows for fine correction of high-frequency distortion at different depths.

[0036] When the compensation is positive, it is added to the output value of the proportional-integral-derivative (PID) controller; when the compensation is negative, its absolute value is subtracted from the output value. Because distortion in the low-speed range causes the actual flow rate to be lower than expected, the controller output needs to be increased, hence the compensation is positive. Conversely, distortion in the high-speed range causes the actual flow rate to be higher than expected, requiring the controller output to be reduced, hence the compensation is negative. For example, if the compensation calculated in the low-speed range is 0.5, it is directly added to the output value of the PID controller to give the inverter a higher frequency command. If the compensation calculated in the high-speed range is -0.3, then 0.3 is subtracted from the output value to avoid pressure overshoot.

[0037] The first, second, third, fourth, and fifth proportional coefficients are calibrated in advance according to the following process. The purpose of these coefficients is to ensure that the square, cube, fourth power, sine, and exponential functions in the piecewise compensation function can accurately offset the actual nonlinear distortion of the frequency converter in different frequency ranges. Therefore, the calibration process must be based on actual field measurement data and cannot be arbitrarily assumed.

[0038] First, within the no-load to full-load range of the inverter compressor, select multiple test frequency points and measure the actual output flow rate at each test frequency point. For example, for an air compressor with a rated frequency of 50 Hz, start from 5 Hz and take a test point every 5 Hz until 50 Hz, thus obtaining 10 test frequency points. During measurement, keep the outlet valve opening of the pipeline constant and allow the air compressor to run stably for at least one minute at each test frequency point. Then, read the values ​​from the vortex flow meter or thermal flow meter. This method provides more reliable data.

[0039] For each test frequency point, the theoretical linear flow rate corresponding to that frequency point is calculated based on the direct proportional relationship between frequency and flow rate. Specifically, the test frequency point is divided by the rated frequency of the inverter host and then multiplied by the rated flow rate. Since, ideally, the output flow rate of the air compressor is directly proportional to the operating frequency, and the rated flow rate is output at the rated frequency, the theoretical linear flow rate is equal to the test frequency divided by the rated frequency and then multiplied by the rated flow rate. For example, if the rated frequency is 50 Hz and the rated flow rate is 10 cubic meters per minute, then the theoretical linear flow rate corresponding to a test frequency of 20 Hz is 20 divided by 50 and then multiplied by 10, which equals 4 cubic meters per minute. The theoretical linear flow rate corresponding to a test frequency of 45 Hz is 45 divided by 50 and then multiplied by 10, which equals 9 cubic meters per minute.

[0040] The nonlinear distortion rate at each test frequency point is obtained by dividing the difference between the actual output flow rate and the theoretical linear flow rate by the theoretical linear flow rate. The distortion rate can be positive or negative; it is positive when the actual flow rate is greater than the theoretical linear flow rate and negative when it is less. For example, at 20 Hz, the actual measured flow rate is 3.6 cubic meters per minute, while the theoretical value is 4. The difference is -0.4. Dividing -0.4 by 4 gives -0.1, or a distortion rate of -10%. At 45 Hz, the actual flow rate is 9.5 cubic meters per minute, while the theoretical value is 9. The difference is +0.5. Dividing 0.5 by 9 gives approximately +0.0556, or a distortion rate of approximately +5.56%.

[0041] The distortion rate curves from zero frequency to a first frequency threshold are fitted with frequency to obtain the first, second, and third proportionality coefficients. Specifically, the distortion rates of all test points between zero and the first frequency threshold are plotted as a curve in frequency order. Then, the distortion rates within each of the previously defined first, second, and third frequency sub-intervals are averaged. The first proportionality coefficient corresponds to the average distortion rate within the first frequency sub-interval, the second proportionality coefficient corresponds to the average distortion rate within the second frequency sub-interval, and the third proportionality coefficient corresponds to the average distortion rate within the third frequency sub-interval. For example, if the first frequency threshold is set to 25 Hz, then the range from zero to 25 Hz is divided into three equal sub-intervals: 0 to 8.33 Hz, 8.33 to 16.67 Hz, and 16.67 to 25 Hz. If the measured average distortion rate in the 0 to 8.33 Hz range is -0.05, then the first proportionality coefficient is set to 0.05. If the average distortion rate in the 8.33 to 16.67 Hz range is -0.12, then the second proportionality coefficient is set to 0.12. If the average distortion rate in the 16.67 to 25 Hz range is -0.08, then the third proportionality coefficient is set to 0.08. Positive numbers are used here because the direction has already been handled using a power function and sign rules in the compensation calculation formula; the proportionality coefficient only indicates the magnitude of the distortion intensity.

[0042] The distortion rate curves within the range from the second frequency threshold to the maximum operating frequency are fitted to obtain the fourth and fifth proportionality coefficients. Similarly, the average distortion rates within the fourth and fifth frequency sub-intervals are taken. The fourth proportionality coefficient corresponds to the average distortion rate within the fourth frequency sub-interval, and the fifth proportionality coefficient corresponds to the average distortion rate within the fifth frequency sub-interval. For example, if the second frequency threshold is set to 45 Hz and the maximum operating frequency is 50 Hz, then 45 to 50 Hz is equally divided into two sub-intervals: 45 to 47.5 Hz and 47.5 to 50 Hz. If the measured average distortion rate within the 45 to 47.5 Hz interval is +0.02, the fourth proportionality coefficient is set to 0.02. If the average distortion rate within the 47.5 to 50 Hz interval is +0.07, the fifth proportionality coefficient is set to 0.07.

[0043] The proportional coefficients calibrated in this way can accurately reflect the strength of actual nonlinear distortion, making the compensation calculation very close to real-world requirements. If the measured distortion rate changes drastically with frequency, the maximum value or weighted average of the distortion rate within a sub-interval can be used instead of the average value. However, to maintain system stability, the average value is generally sufficient. These proportional coefficients are usually calibrated once before the air compressor leaves the factory and written into the controller. If the inverter host ages or the motor is replaced after long-term operation, the above calibration process can be repeated to update the coefficients. The more test frequency points there are, the more accurate the fitted coefficients will be; usually, 10 to 20 points are sufficient. In terms of numerical range, the first to fifth proportional coefficients generally fall between 0 and 0.5, because the actual nonlinear distortion rate rarely exceeds 50%. Finally, these coefficients are stored in the controller's parameter table for direct use when calculating compensation in real time online.

[0044] Generating a virtual lock signal refers to: A pre-established loading / unloading event queue is used to record in real time the timing of the unloading action of each slave unit, as well as the air supply pressure values ​​within one sampling period before and after unloading. The sampling period is typically 0.1 to 0.2 seconds, because the air compressor unloading valve's action time is approximately 0.3 seconds, and one sampling period is sufficient to capture the comparison of pressure changes before and after. For example, if a slave unit performs unloading at 10.0 seconds, the queue will record the moment at 10.0 seconds, along with the air supply pressure at 9.9 seconds and 10.1 seconds. This allows for accurate subsequent determination of whether a pressure rebound occurred after unloading.

[0045] Once any slave device completes its unloading action, the system continuously collects the gas supply pressure value after unloading and calculates the maximum pressure increment within a first preset time period after unloading. This first preset time period is typically 0.5 to 1 second, as pressure rebound due to pipeline cavity effects usually occurs between 0.3 and 0.8 seconds after unloading. The maximum pressure increment is calculated by finding the highest pressure value within the first preset time period and subtracting the initial pressure value immediately after unloading. If this maximum pressure increment exceeds a pre-set rebound threshold, the system determines that an instantaneous pressure rebound has occurred and marks this slave device as the first slave. The rebound threshold is typically set between 0.01 and 0.02 MPa, because after normal unloading, the pressure will only decrease slightly or stabilize; only gas expansion within the pipeline will cause a sudden increase exceeding 0.01 MPa. For example, if the pressure of a slave device rises from 0.73 MPa to 0.75 MPa within 0.6 seconds after unloading, an increase of 0.02 MPa, which exceeds the threshold of 0.015 MPa, then the system will record this slave device as the first slave device.

[0046] Iterate through all slave devices that meet the unloading condition based on the second pressure threshold, excluding the already marked first slave device. The remaining slave devices constitute the candidate slave device set. Meeting the unloading condition based on the second pressure threshold means that the current gas supply pressure is greater than or equal to the second pressure threshold (e.g., 0.7665 MPa). For example, if there are three slave devices A, B, and C, and their pressures all exceed 0.7665 MPa, with A being the first slave device, then the candidate slave device set would include B and C.

[0047] The system dynamically determines a number of slave devices to be disabled based on the maximum pressure increment and the deviation between the current supply pressure and the target pressure. A larger maximum pressure increment indicates a more severe rebound, requiring more slave devices to be disabled. Similarly, a greater deviation between the current supply pressure and the target pressure indicates a further deviation from the normal system, also necessitating a larger number of slave devices to be disabled. The specific calculation method will be disclosed separately later; here, only the two key parameters it relies on are explained.

[0048] Select a number of slave devices from the candidate slave device set according to their priority, and mark each selected slave device as the second slave device. The slave device priority can be arranged in ascending order of cumulative runtime or according to a fixed device number. For example, if there are two slave devices B and C in the candidate slave device set, and the number of slave devices to be banned is 1, with B having priority over C, then only B will be selected and marked as the second slave device. If the number of slave devices to be banned is 2, then both B and C will be selected and marked as the second slave device.

[0049] A virtual lock signal is generated for each second slave device. This virtual lock signal includes the identification of the corresponding second slave device (such as device number) and a lock time window. The duration of the lock time window is a pre-set second preset duration, typically 2 to 3 seconds. This duration cannot be too short, otherwise it will not effectively block the chain reaction; nor can it be too long, otherwise normal regulation will fail. Numerous experiments have shown that 2 seconds can cover a complete pressure drop process after a pressure rebound.

[0050] During the locked time window, the system prohibits all second slave units from responding to any unloading commands based on the second pressure threshold. This means that even if the current gas supply pressure is still higher than the second pressure threshold, these locked slave units will not perform an unloading action. For example, if second slave unit B is locked for 2 seconds, and the pressure remains at 0.77 MPa during these 2 seconds, the controller will not send an unloading command to B.

[0051] After the lockout time window expires, the system automatically clears all virtual lockout signals and restores the normal unloading function of all secondary slave devices. This avoids the risk of a chain reaction of failures without permanently altering the control logic of the slave devices. For example, if the virtual lockout signal of B is cleared after 2 seconds, and the pressure is still higher than the second pressure threshold, B can unload normally. The entire process is completed automatically without manual intervention.

[0052] The specific process for determining the prohibited quantity is as follows: Pre-set a baseline rebound step size, a baseline deviation step size, and a scaling factor. The baseline rebound step size is determined based on the pipeline cavity time constant and the pressure sensor resolution, typically using the ratio of the pipeline pressure rebound peak to the unloading valve's actuation time. For example, when the pipeline volume is 5 cubic meters and the unloading valve closing time is 0.3 seconds, the measured maximum rebound increment is 0.015 MPa; therefore, the baseline rebound step size is 0.005 MPa (one-third of the rebound increment). The baseline deviation step size is set according to the pressure fluctuation range allowed by the production process, generally between one-tenth and one-fifth of the difference between the target pressure value and the minimum allowable pressure. For example, if the target pressure is 0.70 MPa and the minimum allowable pressure is 0.60 MPa, the difference is 0.10 MPa, and the baseline deviation step size can be 0.01 MPa. The proportional factor is determined based on the ratio of the gas supply capacity of a single slave unit to the total demand. When the gas supply capacity of a single slave unit accounts for 20% of the total demand, the proportional factor is 0.8; when it accounts for 10% of the total demand, the proportional factor is 0.5; and when it accounts for 30% of the total demand, the proportional factor is 1.2.

[0053] These parameters can all be calibrated on-site through more than three load change tests. For example, a 0.05 MPa step pressure drop can be created on the main outlet pipe of the air compressor station, and the pressure rebound and deviation values ​​can be recorded to deduce the appropriate step size and scaling factor. The specific numerical ranges are as follows: the reference rebound step size is usually between 0.003 and 0.008 MPa, the reference deviation step size is between 0.008 and 0.015 MPa, and the scaling factor is between 0.5 and 1.2. For example, a reference rebound step size of 0.005 MPa, a reference deviation step size of 0.01 MPa, and a scaling factor of 0.8 can be used.

[0054] First, calculate the first ratio of the maximum pressure increment divided by the baseline rebound step size. Simultaneously, calculate the second ratio of the absolute value of the deviation between the current supply pressure and the target pressure divided by the baseline deviation step size. Then, take the larger of the first and second ratios and multiply this larger value by a scaling factor to obtain the initial prohibition quantity. This is done to ensure that the final prohibition quantity can handle both the severity of pressure rebounds and the severity of pressure deviations from normal values, and that taking the larger value prioritizes more severe conditions. For example, if the maximum pressure increment is 0.02 MPa, the baseline rebound step size is 0.005 MPa, and the first ratio is 4; if the deviation between the current pressure and the target pressure is 0.03 MPa, the baseline deviation step size is 0.01 MPa, and the second ratio is 3. The larger value is 4, which, multiplied by a scaling factor of 0.8, yields an initial prohibition quantity of 3.2.

[0055] The initial prohibition quantity is rounded up to the nearest integer, meaning that regardless of the decimal part, it is always rounded up to the nearest whole number. For example, 3.2 rounded up results in 4, and 2.0 rounded up results in 2. This process avoids insufficient prohibition quantities due to decimal truncation, ensuring that at least the most dangerous units can be suppressed.

[0056] The rounded-up number of prohibited units is limited to the range of the preset maximum prohibited unit, and this is used as the final prohibited unit number. The preset maximum prohibited unit number is usually set to the total number of slave units in the candidate set or a small value (e.g., 3 units). The purpose is to prevent prohibiting too many slave units from causing an excessive decrease in gas supply capacity. For example, if the rounding-up result is 4, but the maximum prohibited unit number is 3, then the final prohibited unit number is 3. If the rounding-up result is 2, and the maximum prohibited unit number is 3, then the prohibited unit number is 2. This ensures the blocking effect while avoiding excessive suppression. For example, if there are 5 slave units on site that meet the unloading conditions, and after excluding the first slave unit, there are 4 remaining, and the maximum prohibited unit number is set to 3, the calculated initial prohibited unit number is 2.1. Rounding up gives 3, which does not exceed the maximum prohibited unit number, so the prohibited unit number is 3 units. If the initial prohibited unit number is 3.8, rounding up gives 4, which exceeds the maximum prohibited unit number of 3, then the final prohibited unit number is 3 units.

[0057] The re-optimization of the forced trigger scheduling strategy refers to: The re-optimization includes correcting the target pressure setpoint of the proportional-integral-derivative controller, reallocating the master-slave ratio, and adjusting the baseline values ​​of the first and second pressure thresholds. Specifically, the target pressure is first increased by the net cumulative error multiplied by a first preset value. The net cumulative error, as previously described, represents the algebraic sum of pressure deviations over a period of time; a positive value indicates a persistently high pressure, while a negative value indicates a persistently low pressure. The first preset value is a dimensionless decimal, typically between 0.001 and 0.01, for example, 0.005. The setting is based on the principle that each adjustment should not be too large, otherwise it will cause pressure oscillations, nor too small, otherwise convergence will be too slow. For example, if the net cumulative error is 20 MPa / s and the first preset value is 0.005, then the target pressure will increase by 0.1 MPa, from 0.70 MPa to 0.80 MPa. This raises the setpoint, thereby reducing frequent unloading.

[0058] Calculate the ratio of the arithmetic mean of the running time of each slave device to the fourth power of the root mean difference, then multiply it by the ratio of the median to the maximum absolute deviation, and finally multiply by a preset conversion value to obtain the correction value, where the maximum absolute deviation is the maximum absolute value of the difference between the running time of each slave device and the median.

[0059] For example, the running times of three slave machines are 502 hours, 500 hours, and 498 hours, respectively, with an arithmetic mean of 500 hours. The differences between each duration and the mean are 2 hours, 0 hours, and -2 hours, respectively, and raised to the fourth power are 16, 0, and 16. Summing these values ​​gives 32, dividing by 3 gives approximately 10.67, and taking the fourth root gives approximately 1.8. Dividing the arithmetic mean by this value gives 500 divided by 1.8, which is approximately 278. The median is 500 hours, and the largest absolute deviation is 2 hours (502 minus 500). Dividing the median by the largest absolute deviation gives 500 divided by 2, which equals 250. The product of these two ratios, 278 multiplied by 250, equals 69,500. Multiplying this by a preset conversion value (set to 0.00001) gives 0.00695. We take an approximate 0.007 as the correction value.

[0060] Multiplying the second preset value (0.001) by the correction value of 0.007, and then multiplying by the cumulative error (set to 50 MPa / s), we obtain the adjusted pressure value, which equals 0.001 multiplied by 0.007 multiplied by 50, equaling 0.00035 MPa, or 0.35 kPa. This adjusted pressure value is used to lower the reference value of the first pressure threshold, and the same adjusted pressure value is used to raise the reference value of the second pressure threshold. The original reference values ​​were 0.68 MPa and 0.73 MPa, which are adjusted to 0.67965 MPa and 0.73035 MPa, respectively. The change is minimal, meeting the adaptive correction requirements for slow drift after long-term operation.

[0061] Finally, the adjusted pressure value is used to lower the baseline value of the first pressure threshold, and the same adjusted pressure value is used to raise the baseline value of the second pressure threshold. That is, the baseline value of the first pressure threshold decreases by 0.075 MPa, and the baseline value of the second pressure threshold increases by 0.075 MPa. For example, the original baseline values ​​were 0.68 and 0.73, which become 0.605 and 0.805 MPa after adjustment. This expands the dead zone, allowing the system to tolerate greater pressure fluctuations and reducing the number of unit start-ups and shutdowns. If the cumulative error is small, the adjusted pressure value is also small, and the threshold remains almost unchanged. The reason for choosing the fourth root mean difference and the maximum absolute deviation is: the fourth root mean is very sensitive to extreme running times; when a slave unit is not scheduled for a long period (the time is significantly shorter), this value will increase rapidly, thereby reducing the correction value and preventing excessive expansion of the dead zone that would cause the slave unit to be idle for a long time; the maximum absolute deviation reflects the dispersion of the middle 50% of the data, resisting outlier interference. The combination of the two can comprehensively assess the load balancing status. All parameters (first preset value, second preset value, preset conversion value) can be fine-tuned according to the actual response speed during system debugging. The generally recommended range is: first preset value 0.001-0.005, second preset value 0.0005-0.002, and preset conversion value 0.5-2.

[0062] The following is a complete embodiment of the present invention, used to illustrate the operation process of the entire switching control method: I. System Configuration and Initial Parameters: A certain air compressor station has one variable frequency drive (rated frequency 50 Hz, rated flow rate 10 cubic meters per minute) and three power frequency slave units (numbered 1, 2, and 3). The target pressure is set to 0.70 MPa. The preset parameters are as follows: First frequency threshold: 25 Hz; Second frequency threshold: 45 Hz. First pressure threshold reference value: 0.68 MPa (starting slave unit); Second pressure threshold reference value: 0.73 MPa (unloading slave unit). Delay confirmation time reference value: 1 second; Fluctuation identification frequency threshold value: 0.15 Hz; Amplitude threshold value: 0.008 MPa per second; Rebound threshold value: 0.015 MPa; First preset duration: 0.5 seconds; Second preset duration: 2 seconds; Reference rebound step size: 0.005 MPa; Reference deviation step size: 0.01 MPa; Scale factor: 0.8; Maximum prohibited number: three units; First preset value: 0.005; Second preset value: 0.001; Preset conversion value: 1.

[0063] II. Fluctuation Identification Stage: The system continuously collects gas supply pressure at a sampling interval of 0.1 seconds. At a certain moment, the gas consumption suddenly increases from 30 cubic meters per minute to 70 cubic meters per minute, and the pressure fluctuates rapidly between 0.65 and 0.75 MPa, with a period of approximately 5 seconds. A window of data from 128 sampling points (12.8 seconds) prior to the current moment is taken. The maximum pressure within the window is 0.75 MPa, and the minimum is 0.65 MPa, with a difference of 0.10 MPa. Dividing this by 12.8 seconds yields an amplitude change rate of 0.0078 MPa per second. A Fast Fourier Transform (FFT) is performed on the window pressure signal, revealing that the frequency component with the largest amplitude is 0.2 Hz. 0.0078 is less than 0.008, failing to meet the "greater than amplitude threshold" requirement. Therefore, the current condition does not meet the second level, and the first level (non-high-frequency intermittent fluctuation) is output.

[0064] If the gas fluctuation period is four seconds, the main frequency is 0.25 Hz, and the pressure changes rapidly between 0.64 and 0.76 Hz, with a maximum pressure of 0.76 Hz and a minimum of 0.64 Hz within the window, the difference is 0.12 Hz. Dividing this by 12.8 gives 0.0094 MPa per second, which exceeds 0.008 Hz. In this case, the main frequency of 0.25 Hz is greater than 0.15 Hz, and the amplitude change rate of 0.0094 Hz is greater than 0.008 Hz. Therefore, the output is at the second level (high-frequency intermittent fluctuation state).

[0065] III. Adaptive Dead-Zone Adjustment: When outputting the second level, the system performs dynamic adjustment. The first pressure threshold decreases from the baseline value of 0.68 MPa by a first preset percentage (5%), becoming 0.646 MPa. The second pressure threshold increases from 0.73 MPa by a second preset percentage (also 5%), becoming 0.7665 MPa. The delay confirmation time increases from one second by 0.5 times, becoming 1.5 seconds. This makes it easier for the slave device to start and more difficult to unload, and allows for a longer waiting time before judgment, avoiding frequent actions.

[0066] IV. Nonlinear Compensation: At this time, the operating frequency of the inverter is 22 Hz (less than 25 Hz). The frequency range from 0 to 25 Hz is divided into three sub-intervals: 0 to 8.33 Hz, 8.33 to 16.67 Hz, and 16.67 to 25 Hz. 22 Hz falls within the third sub-interval (16.67 to 25 Hz). The difference between the operating frequency and the first frequency threshold is 3 Hz (25 minus 22). The compensation amount is equal to the third proportional coefficient multiplied by the fourth power of the difference. Assuming the third proportional coefficient is 0.01 obtained through calibration, the compensation amount is equal to 0.01 multiplied by 3 to the fourth power (81), which equals 0.81. This compensation amount is positive and is superimposed on the output value of the proportional-integral-derivative controller, increasing the actual frequency command of the inverter by 0.81 Hz, reaching approximately 22.81 Hz, thereby compensating for insufficient flow in the low-speed zone.

[0067] V. Chain Disturbance Interruption: The pressure gradually rises to 0.77 MPa, exceeding the second pressure threshold of 0.7665 MPa. Slave 1 meets the unloading condition and executes unloading. Within 0.4 seconds after unloading, the pressure rebounds from 0.77 to 0.785 MPa (an increment of 0.015 MPa), equal to the rebound threshold. This indicates an instantaneous pressure rebound characteristic, and slave 1 is marked as the first slave. At this time, slaves 2 and 3 also meet the unloading condition because their pressure of 0.785 is greater than 0.7665. After excluding the first slave, the candidate slave set includes slaves 2 and 3. The maximum pressure increment is 0.015 MPa, and the absolute value of the deviation between the current pressure and the target pressure is 0.785 minus 0.70, which equals 0.085 MPa.

[0068] Calculate the first ratio: 0.015 divided by 0.005 equals 3; the second ratio: 0.085 divided by 0.01 equals 8.5. Take the larger value, 8.5, and multiply it by the scaling factor 0.8 to get 6.8. Round up to get 7. Set the maximum number of prohibited units to three, and the limited number of prohibited units is three. According to priority order (e.g., shorter running time takes precedence), select three units from the candidate set (i.e., units two and three, but only two are actually selected), and mark both of them as the second slave. Generate a virtual lock signal for each unit, with a lock time window of two seconds. Within two seconds, units two and three are prohibited from unloading. After two seconds, the pressure naturally drops back to 0.74 MPa, the unloading condition disappears, and the virtual lock signal is automatically cleared. This avoids the gas supply cliff caused by units two and three unloading simultaneously after unit one unloads.

[0069] VI. Pressure Error Feedback Integration and Re-optimization: Assume the system has run for eight hours, continuously calculating the difference between the actual pressure and the target pressure. The net cumulative error (algebraic sum) is +30 MPa / s, and the cumulative error (absolute value integral) is 120 MPa / s. If the cumulative error exceeds the preset error threshold (set to 80 MPa / s), re-optimization is triggered. First, the target pressure is corrected: 0.70 MPa is increased by 30 times the net cumulative error multiplied by the first preset value of 0.005, i.e., an increase of 0.15 MPa, resulting in a new target pressure of 0.85 MPa. The running durations of each slave are 500 hours for slave 1, 480 hours for slave 2, and 520 hours for slave 3, with an arithmetic mean of 500 hours. The fourth root mean difference is calculated: the difference between each duration and the average is 0, -20, and +20, respectively; the fourth root is 0, 160,000, and 160,000, respectively, averaging approximately 100,000.67, and taking the fourth root yields approximately 17.8. The arithmetic mean is divided by this value: 500 divided by 17.8 is approximately 28. The median is 500 hours, and the maximum absolute deviation is 20 hours (520 minus 500). The median divided by the maximum absolute deviation is 500 divided by 20, which equals 25. The product 28 multiplied by 25 equals 700. Multiplying this by the preset conversion value 1 gives the correction value of 700.

[0070] Calculate the adjusted pressure value: Multiply the second preset value of 0.001 by the correction value of 700, then multiply by the cumulative error of 120, resulting in 84 MPa. This value is too large, indicating it is unreasonable, because the preset conversion value should be adjusted (the actual correction value should be between 0.5 and 2). For illustration, the example parameters are adjusted to make the correction value reasonable, assuming a more uniform distribution of running time: 500 for No. 1, 501 for No. 2, and 502 for No. 3, with an average of 501, a fourth-power root mean difference of approximately 0.8, and a ratio of 626; the median is 501, the maximum absolute deviation is 1, and the ratio is 501. The product is 31,000, which is still very large. This shows that the preset conversion value in the original design should be very small (e.g., 0.0001). In fact, the main function of the correction value is fine-tuning; therefore, if the preset conversion value is correctly set to 0.001, the correction value will be approximately 0.7. The pressure value is adjusted to 0.001 multiplied by 0.7 multiplied by 120, which equals 0.084 MPa. Then, the first pressure threshold reference value is reduced by 0.084 MPa, from 0.68 to 0.596 MPa. At the same time, the second pressure threshold reference value is increased by the same value, from 0.73 to 0.814 MPa. This expands the dead zone and adjusts the target pressure, making subsequent control more adaptable to the operating conditions after long-term operation.

[0071] VII. Closed-Loop Feedback: The re-optimized parameters (new target pressure, new first and second pressure threshold benchmarks, and adjusted slave priority order) are used for subsequent control. The new target pressure is 0.735 MPa (assuming a small net cumulative error), the threshold benchmarks are updated, and the system returns to the fluctuation identification phase to continue monitoring the pressure signal.

[0072] The above algorithms or formulas are all dimensionless and numerical calculations, and the results are obtained by software simulation based on a large amount of collected data to obtain the most recent real-world results. The preset parameters are set by those skilled in the art according to the actual situation.

[0073] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0074] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0075] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0076] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for switching operating modes of an air compressor, characterized in that, Includes the following steps: The pressure signal in the gas supply pipeline is collected in real time. Frequency domain and time domain analysis is performed on the pressure signal to calculate the rate of change of pressure fluctuation amplitude and the frequency components of fluctuation. Based on the rate of change of pressure fluctuation amplitude and the frequency components of fluctuation, the fluctuation state of the current gas load is determined and the fluctuation level parameter is output. The preset pressure switching dead zone threshold in the linkage control of multiple air compressors is dynamically adjusted according to the fluctuation level parameter, and the delay confirmation time is also adjusted. The operating frequency of the current inverter host is obtained. When the operating frequency is less than the first frequency threshold or greater than the second frequency threshold, the compensation amount is calculated according to the piecewise compensation function and the compensation amount is superimposed on the output value of the proportional-integral-derivative controller to correct the deviation between the gas supply pressure and the target pressure. The first frequency threshold is less than the second frequency threshold. Monitor the loading and unloading action event sequence of all air compressor units and the instantaneous pressure rebound characteristics after unloading. When the air supply pressure rebounds after the first slave unit is unloaded and the second slave unit meets the condition of the unloading threshold based on the preset pressure switching dead zone threshold, generate a virtual lock signal and temporarily prohibit all determined second slave units from performing unloading actions within the preset time window. The error between the actual gas supply pressure and the target pressure is continuously calculated, and the absolute value of the error is integrated over time to obtain the cumulative error. At the same time, the error value itself is integrated over time to obtain the net cumulative error. When the cumulative error exceeds the preset error threshold, the re-optimization of the scheduling strategy is forcibly triggered.

2. The method for switching operating modes of an air compressor according to claim 1, characterized in that, The fluctuation level parameters include a first level representing non-high-frequency intermittent fluctuation states and a second level representing high-frequency intermittent fluctuation states.

3. The method for switching operating modes of an air compressor according to claim 2, characterized in that, The specific process for determining the fluctuation state of the current gas load based on the rate of change of pressure fluctuation amplitude and the frequency components of the fluctuation includes: Pressure time series is obtained by continuously collecting pressure signals in the gas supply pipeline at preset sampling intervals. Pressure data within a sliding window before the current moment is taken. The length of the sliding window is a preset number of sampling points. Calculate the difference between the maximum and minimum pressure values ​​within the sliding window, divide it by the total duration of the sliding window, and obtain the rate of change of pressure fluctuation amplitude. Perform a fast Fourier transform on the pressure signal within the sliding window and extract the frequency component with the largest amplitude as the main frequency. Compare the main frequency with the preset frequency threshold value, and compare the pressure fluctuation amplitude change rate with the preset amplitude threshold value. When the main frequency is greater than the frequency threshold and the rate of change of pressure fluctuation amplitude is greater than the amplitude threshold, the output is the second level fluctuation level parameter representing the high-frequency intermittent fluctuation state; otherwise, the output is the first level fluctuation level parameter representing the non-high-frequency intermittent fluctuation state.

4. The method for switching operating modes of an air compressor according to claim 1, characterized in that, The pressure switching dead zone threshold includes a first pressure threshold for starting the slave and a second pressure threshold for unloading the slave.

5. The method for switching operating modes of an air compressor according to claim 4, characterized in that, The specific process of dynamically adjusting the pressure switching dead zone threshold in the multi-air compressor linkage control based on the fluctuation level parameter, and simultaneously adjusting the delay confirmation time, includes: The baseline values ​​for the first pressure threshold, the second pressure threshold, and the delayed confirmation time are preset. When the fluctuation level parameter is the first level representing a non-high-frequency intermittent fluctuation state, the first pressure threshold and the second pressure threshold are kept as their respective baseline values, and the delayed confirmation time is set as the baseline value. When the fluctuation level parameter is the second level representing a high-frequency intermittent fluctuation state, the first pressure threshold is reduced from the reference value by a first preset ratio, the second pressure threshold is increased from the reference value by a second preset ratio, and the delay confirmation time is increased from the reference value by a preset multiple.

6. The method for switching operating modes of an air compressor according to claim 5, characterized in that, The specific process of calculating the compensation amount based on the piecewise compensation function includes: The interval from zero frequency to the first frequency threshold is pre-divided into a first sub-interval, a second sub-interval, and a third sub-interval, and the interval from the second frequency threshold to the maximum operating frequency is pre-divided into a fourth sub-interval and a fifth sub-interval. Obtain the current operating frequency of the inverter host. When the operating frequency is greater than or equal to the first frequency threshold and less than or equal to the second frequency threshold, the compensation amount is zero. When the operating frequency is less than the first frequency threshold, the compensation amount is calculated according to the frequency sub-interval that the operating frequency falls into: When the frequency falls into the first sub-interval, the compensation amount is equal to the first proportional coefficient multiplied by the square of the difference between the operating frequency and the first frequency threshold. When the value falls into the second subinterval, the compensation amount is equal to the second proportional coefficient multiplied by the cube of the difference; when the value falls into the third subinterval, the compensation amount is equal to the third proportional coefficient multiplied by the fourth power of the difference. When the operating frequency is greater than the second frequency threshold, the compensation amount is calculated according to the frequency sub-intervals into which the operating frequency falls: When the frequency falls into the fourth sub-interval, the compensation amount is equal to the fourth proportional coefficient multiplied by the sine value of the difference between the operating frequency and the second frequency threshold. When the value falls into the fifth subinterval, the compensation amount is equal to the fifth proportional coefficient multiplied by the exponential function value of the difference. The exponential function value is equal to the difference of the natural constant raised to the power of one. When the compensation is positive, it is added to the output value of the proportional-integral-derivative controller; when the compensation is negative, its absolute value is subtracted from the output value.

7. The method for switching operating modes of an air compressor according to claim 6, characterized in that, The first, second, third, fourth, and fifth scaling factors are pre-calibrated according to the following process: Within the range of no-load to full-load of the inverter host, select multiple test frequency points and measure the actual output flow at each test frequency point; For each test frequency point, the theoretical linear flow rate corresponding to that frequency point is calculated based on the direct proportional relationship between frequency and flow rate. Specifically, the test frequency point is divided by the rated frequency of the inverter host and then multiplied by the rated flow rate. Divide the difference between the actual output flow rate and the theoretical linear flow rate at each test frequency point by the theoretical linear flow rate to obtain the nonlinear distortion rate at that frequency point. The curves of distortion rate versus frequency within the range from zero frequency to the first frequency threshold are fitted to obtain the first proportional coefficient, the second proportional coefficient, and the third proportional coefficient, respectively. The first proportional coefficient corresponds to the average distortion rate within the first frequency sub-interval, the second proportional coefficient corresponds to the average distortion rate within the second frequency sub-interval, and the third proportional coefficient corresponds to the average distortion rate within the third frequency sub-interval. The distortion rate curves within the range from the second frequency threshold to the maximum operating frequency are fitted to obtain the fourth and fifth proportional coefficients, respectively. The fourth proportional coefficient corresponds to the average distortion rate within the fourth frequency sub-interval, and the fifth proportional coefficient corresponds to the average distortion rate within the fifth frequency sub-interval.

8. The method for switching operating modes of an air compressor according to claim 7, characterized in that, Generating a virtual lock signal refers to: A pre-established loading / unloading action event queue is used to record the time of unloading action of each slave unit and the gas supply pressure value within one sampling cycle before and after unloading in real time. When any slave device completes the unloading action, the gas supply pressure value after unloading is continuously collected in the event queue. The maximum pressure increment within the first preset time after unloading is calculated. If the maximum pressure increment exceeds the preset rebound threshold, it is determined that an instantaneous pressure rebound feature has occurred, and the slave device is marked as the first slave device. Iterate through all slave devices that meet the unloading condition based on the second pressure threshold, and exclude the first slave device to obtain the set of candidate slave devices; A prohibited quantity is dynamically determined based on the maximum pressure increment and the deviation between the current gas supply pressure and the target pressure. Select a prohibited number of slave devices from the set of candidate slave devices according to their priority, and mark each selected slave device as the second slave device. A virtual locking signal is generated for each second slave device. The virtual locking signal contains the identity of the corresponding second slave device and a locking time window. The duration of the locking time window is a preset second preset duration. During the locked time window, all second slave devices are prohibited from responding to any unload commands based on the second pressure threshold; After the lockout time window ends, all virtual lockout signals will be automatically cleared, and the normal unloading function of all second slave devices will be restored.

9. A method for switching operating modes of an air compressor according to claim 8, characterized in that, The specific process for determining the prohibited quantity is as follows: Pre-set a reference bounce step size, a reference deviation step size, and a scaling factor; The first ratio of the maximum pressure increment to the baseline rebound step size is calculated, and the second ratio of the absolute value of the deviation between the current gas supply pressure and the target pressure to the baseline deviation step size is calculated. The larger of the first and second ratios is taken, and the larger value is multiplied by the scaling factor to obtain the initial prohibition quantity. The initial number of prohibited items is rounded up to obtain the rounded number of prohibited items. The number of prohibited items after rounding is limited to the range of the preset maximum number of prohibited items, and is used as the number of prohibited items.

10. A method for switching operating modes of an air compressor according to claim 9, characterized in that, The re-optimization of the forced trigger scheduling strategy refers to: The re-optimization includes correcting the target pressure setpoint of the proportional-integral-derivative controller, reallocating the master-slave ratio, and adjusting the baseline values ​​of the first and second pressure thresholds, specifically: Increase the target pressure by the net cumulative error multiplied by the first preset value; Calculate the ratio of the arithmetic mean of the running time of each slave to the fourth root mean difference, then multiply it by the ratio of the median to the maximum absolute deviation, and finally multiply it by a preset conversion value to obtain the correction value, where the maximum absolute deviation is the maximum value of the absolute value of the difference between the running time of each slave and the median. Multiply the second preset value by the correction value and then by the cumulative error to obtain an adjustment pressure value; The adjustment pressure value is used to lower the reference value of the first pressure threshold, and the same adjustment pressure value is used to raise the reference value of the second pressure threshold.

Images