Variable-frequency screw air compression system control method based on data-driven PID (Proportion Integration Differentiation) control

Abstract

The invention discloses a variable-frequency screw air compression system control method based on data-driven PID control, and the method comprises the steps: obtaining a pressure value of an air storage tank and a set pressure value, calculating the adjustment amount of the pressure value of the air storage tank through a data-driven PID control algorithm, and determining the rotating speed of an air compressor or the opening degree of an electromagnetic valve according to the adjustment amount; wherein the difference value between the pressure value of the gas storage tank and the set pressure value is used for calculating a control parameter of the data driving PID control algorithm, so that the difference value between the pressure value of the gas storage tank and the set pressure value is reduced; through the data-driven PID control algorithm, the problem that the output pressure of the air compression system is unstable under the condition that the air compression system model is difficult to accurately establish or the model changes along with time is solved, the pressure change requirements under various complex working conditions can be more flexibly met, and the system reliability is improved. Therefore, the stability and efficiency of the whole compressed air supply network are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of energy saving of compressed air systems, and in particular relates to a control method for a variable frequency screw air compressor system based on data driven PID control. Background Art

[0002] There is a problem of energy waste in the working process of screw air compressors. When the output pressure is greater than a certain value, the unloading valve will automatically open, causing the asynchronous motor to idle, which is an uneconomical operating state and causes serious energy waste. Frequency conversion technology can automatically adjust the running speed of the motor according to the actual gas demand to avoid unnecessary energy consumption. PID (proportional-integral-differential) control, as an important algorithm in frequency conversion control, can control the output pressure so that the air compressor can maintain efficient operation under different working conditions, thereby achieving the purpose of energy saving.

[0003] Traditional PID control relies on accurate mathematical models for parameter setting and system control. However, in practical applications, many air compressor system models are difficult to accurately establish or the models change over time. Especially after the PID parameters are fixed, the adaptability to system changes is relatively weak, and it may not be able to respond effectively to certain complex interference situations, thus affecting the stability of the output pressure and causing unstable air supply pressure. Summary of the invention

[0004] The purpose of the present invention is to provide a variable frequency screw air compressor control method based on data driven PID control to accurately control the output pressure and achieve the purpose of constant pressure air supply.

[0005] The present invention adopts the following technical solutions:

[0006] A control method for a variable frequency screw air compressor system based on data driven PID control comprises the following steps:

[0007] Obtaining the pressure value and set pressure value of the gas storage tank, calculating the adjustment amount of the pressure value of the gas storage tank by using a data-driven PID control algorithm, and determining the speed of the air compressor or the opening degree of the solenoid valve according to the adjustment amount;

[0008] The control parameters of the data-driven PID control algorithm are calculated using the difference between the pressure value of the gas storage tank and the set pressure value to reduce the difference between the pressure value of the gas storage tank and the set pressure value.

[0009] Furthermore, the PID control algorithm adopts the PI control algorithm structure.

[0010] Furthermore, the PI control algorithm structure is:

[0011] Δu(t)=K I (t)e(t)-K P(t)Δy(t),

[0012] Among them, u(t) is the adjustment amount of the gas tank pressure value at the current moment, K I (t) and K P (t) is the control parameter of the PI control algorithm, K I (t) is the ratio of the PI control algorithm at the current moment, K P (t) is the integral gain of the PI control algorithm at the current moment, y(t) is the pressure value of the gas tank at the current moment, e(t)=r(t)-y(t) is the difference between the set pressure value at the current moment and the pressure value of the gas tank, r(t) is the set pressure value at the current moment, and t represents the current moment.

[0013] Furthermore, the control parameter calculation process of the PI control algorithm includes:

[0014] The PI control parameters determined based on historical experience are used as the initial database;

[0015] Calculate the distance between the information vector at the current moment and the information vector in the initial database;

[0016] Select k groups of information vectors with the smallest distance from the initial database;

[0017] Using k groups of information vectors, linear weighting is used to calculate the estimated value of the PI control parameter vector at the current moment;

[0018] The steepest descent method is used to optimize the estimated value of the PI control parameter vector at the current moment to obtain a new PI control parameter vector;

[0019] Update the new PI control parameter vector to the database.

[0020] Furthermore, the expression of the initial database is:

[0021] Φ(j)=[H(j),η(j)],j=1,2,…,N(0),

[0022] Where N(0) is the number of information vectors stored in the initial database, H(j) = [r(j+1), r(j), y(j), …, y(jn y +1),u(j-1),…,u(jn u +1)] is an information vector, including r(j+1) is the set pressure value at time j+1, r(j) is the set pressure value at time j, y(j) is the gas tank pressure value at time j, u(j-1) is the adjustment amount of the gas tank pressure value at time j-1, and the information length of the information vector is n y +n u +1, and n y ≥3 and n u ≥2, ny is the output order, n u is the order of the input, η(j)=[K P (j),K I (j)] is the PI control parameter vector at time j, K P (j) is the integral gain of the time-varying controller at time j, K I (j) is the ratio of the time-varying controller at time j, and j represents time j.

[0023] Furthermore, the distance between the information vector at the current moment and the information vector in the initial database is calculated, and the expression is:

[0024]

[0025] Where N(t) is the number of information vectors in the database at the current moment, H(t) is the information vector at the current moment, l represents the lth element, and H l (t) is the lth element of the information vector at the current moment, H l (j) is the lth element of the jth information vector, maxH l (m) is the maximum value of the lth element of all information vectors in the database, minH l (m) is the minimum value of the lth element of all information vectors in the database.

[0026] Furthermore, the expression of the estimated value of the PI control parameter vector at the current moment is:

[0027]

[0028] Among them, ω i represents the weight of the i-th information vector, and η(i) is the i-th PI control parameter vector.

[0029] Furthermore, the expression of the new PI control parameter vector is:

[0030]

[0031] Among them, λ is the learning factor, which is an adjustable parameter; is the performance index, y r (t+1) is the output at time t+1, To control the gradient of the performance index to the PI parameters.

[0032] Furthermore, the weight of the i-th information vector is expressed as:

[0033]

[0034] An electronic device, comprising:

[0035] at least one processor;

[0036] and, a memory communicatively coupled to the at least one processor;

[0037] The memory stores instructions that can be executed by at least one processor, and the instructions are executed by at least one processor so that at least one processor can execute any one of the above-mentioned variable frequency screw air compressor system control methods based on data-driven PID control.

[0038] The beneficial effect of the present invention is that the present invention utilizes a data-driven PID algorithm to adjust the parameters of the data-driven PID control algorithm in real time according to the pressure value of the gas storage tank, thereby regulating the speed of the air compressor or the opening degree of the solenoid valve in real time, accurately controlling the output pressure, and achieving the purpose of constant pressure gas supply. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 The present invention is a flow chart of a PID control system based on data drive;

[0040] Figure 2 The gas load curve diagram of a workshop when the method of the present invention is applied;

[0041] Figure 3 The figure is a curve diagram showing the effect of the method of the present invention when applied to a certain workshop. DETAILED DESCRIPTION

[0042] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0043] The present invention provides a variable frequency screw air compressor system control method based on data-driven PID control. The data-driven PID control does not rely on an accurate mathematical model. It designs a controller by analyzing the input and output data of the system. Therefore, it has better adaptability to systems with unknown or uncertain models. The control parameters can be dynamically adjusted according to the real-time collected data to adapt to changes in the system. This adaptive capability enables the control system to maintain good control performance when facing external environmental changes, system parameter drift, etc. The use of a data-driven PID algorithm can make the anti-interference ability of the screw air compressor stronger, easier to implement and maintain in practical applications, reduce the professional requirements for technicians, and further achieve a stable supply of pressure.

[0044] A control method for a variable frequency screw air compressor system based on data driven PID control comprises the following steps:

[0045] S110, establish an energy efficiency model for variable frequency screw air compressor.

[0046] Based on the performance analysis of various compressors, the performance calculation of the air compressor at the base frequency is used as a reference, and polynomial fitting is used to obtain the performance of the variable frequency air compressor. Under this condition, the volumetric efficiency, suction volume and energy consumption can be expressed as follows:

[0047]

[0048] Where: Q o_fre Indicates the air intake volume under variable frequency regulation (m 3 / min), V suc,ref Indicates the suction volume of the air compressor per minute at rated frequency (m 3 ), V suc_fre Indicates the suction volume per minute of the air compressor under variable frequency regulation (m 3 ), η v Indicates the volumetric efficiency at rated frequency, η v_fre represents the volumetric efficiency under variable frequency regulation, N represents the speed under variable frequency regulation (rad / min), N ref Indicates the speed at rated frequency (rad / min), f ref represents the rated frequency (Hz), f represents the frequency under variable frequency regulation (Hz), a1, a2, a3 represent the fitting coefficients, W fre represents the power (kW) under variable frequency regulation, W represents the power (kW) under rated frequency, b1, b2, b3 represent fitting coefficients;

[0049] The relationship between speed and frequency is as follows:

[0050]

[0051] In the formula, n represents the motor speed (rad / min), f represents the variable frequency motor operating frequency (Hz), P n It represents the number of pole pairs of the variable frequency motor, and s represents the slip rate.

[0052] S120, calculating the pressure P of the gas storage tank according to the inlet and outlet flow rates of the gas storage tank.

[0053] The formula for calculating the pressure of the gas tank is as follows:

[0054]

[0055] In the formula, T represents the tank temperature (K), m represents the gas mass flow rate (kg / min), m i 、m o are respectively the gas mass flow rate at the inlet and outlet of the gas storage tank (kg / min), m i and m o The difference between represents the change in mass per unit time, VM is the molar volume of the gas (m 3 / mol), a and b are Van der Waals constants, which vary with different substances. The Van der Waals constants a and b of air are 0.1358 and 0.0364 respectively, and k is the Boltzmann constant, which is 1.38*10-23 (J / K).

[0056] S130, comparing the air tank pressure P with the set pressure, calculating the adjustment amount of the air tank pressure value using a data-driven PID control algorithm, and determining the speed of the air compressor or the opening degree of the solenoid valve according to the adjustment amount;

[0057] The control parameters of the data-driven PID control algorithm are calculated in real time using the difference between the gas tank pressure value P and the set pressure value to reduce the difference between the gas tank pressure value and the set pressure value.

[0058] The PID control algorithm in the present invention adopts the PI control algorithm structure, and the structure of the PI control algorithm is:

[0059] Δu(t)=K I (t)e(t)-K P (t)Δy(t),

[0060] Among them, u(t) is the adjustment amount of the gas tank pressure value at the current moment, K I (t) and K P (t) is the control parameter of the PI control algorithm, K I (t) is the ratio of the PI control algorithm at the current moment, K P (t) is the integral gain of the PI control algorithm at the current moment, y(t) is the pressure value of the gas tank at the current moment, e(t)=r(t)-y(t) is the difference between the set pressure value at the current moment and the pressure value of the gas tank, r(t) is the set pressure value at the current moment, and t represents the current moment.

[0061] The PI control algorithm combines proportional and integral control, which can not only respond to deviations quickly but also eliminate steady-state errors. When there is a difference between the air tank pressure value and the set pressure value, control and adjustment are performed quickly, which reduces the reaction time and improves the working efficiency of the air compression system.

[0062] Based on the above foundation, the parameter optimization of the data-driven PID control algorithm can be divided into the steps of accumulating historical information vectors during operation, calculating the estimated values ​​of the PID control algorithm parameters, and optimizing and updating the parameters based on the target performance indicators.

[0063] like Figure 1 As shown, the control parameter calculation steps of the PI control algorithm include:

[0064] S131, the PI control parameters determined based on historical experience are used as the initial database, and the expression of the initial database is:

[0065] Φ(j)=[H(j),η(j)],j=1,2,…,N(0),

[0066] Where N(0) is the number of information vectors stored in the initial database, H(j) = [r(j+1), r(j), y(j), …, y(jn y +1),u(j-1),…,u(jn u +1)] is an information vector, including r(j+1) is the set pressure value at time j+1, r(j) is the set pressure value at time j, y(j) is the gas tank pressure value at time j, u(j-1) is the adjustment amount of the gas tank pressure value at time j-1, and the information length of the information vector is n y +n u +1, and n y ≥3 and n u ≥2, n y is the output order, n u is the order of the input, η(j)=[K P (j),K I (j)] is the PI control parameter vector at time j, K P (j) is the integral gain of the time-varying controller at time j, K I (j) is the ratio of the time-varying controller at time j, and j represents time j.

[0067] S132, calculating the distance between the information vector at the current moment and the information vector in the initial database, expressed as:

[0068]

[0069] Where N(t) is the number of information vectors in the database at the current moment, H(t) is the information vector at the current moment, l represents the lth element, and H l (t) is the lth element of the information vector at the current moment, H l (j) is the lth element of the jth information vector, maxH l (m) is the maximum value of the lth element of all information vectors in the database, minH l (m) is the minimum value of the lth element of all information vectors in the database.

[0070] S133, select k groups of information vectors with the smallest distance from the initial database.

[0071] S134, using k groups of information vectors, linear weighting is used to calculate the estimated value of the PI control parameter vector at the current moment, expressed as:

[0072]

[0073] Among them, η(i) is the i-th PI control parameter vector, ω i represents the weight of the i-th information vector, expressed as:

[0074]

[0075] S135, using the steepest descent method to optimize the estimated value of the PI control parameter vector at the current moment, to obtain a new PI control parameter vector, which is expressed as:

[0076]

[0077] Among them, λ is the learning factor, which is an adjustable parameter; is the performance index, y r (t+1) is the output of the reference model at time t+1, To control the gradient of the performance index to the PI parameter, that is:

[0078]

[0079] in, is the Jacobian matrix of the system, which represents the sensitivity of the controlled object to the change of the control quantity, and is generally approximated by the following formula:

[0080]

[0081] S136, the new PI control parameter vector [H(t),η new (t)]Update to the database.

[0082] From the above analysis, we can know that the parameters that need to be selected are the reference model, the initial parameters of the PI controller, the learning factor and the information length. The reference model selection needs to select a model that is as close to the characteristics of the controlled object as possible. If the selected model responds faster than the controlled object, the system response will be accelerated. An extreme case: the reference model selection is consistent with the set value, that is, it is hoped that the closed-loop system has a faster system dynamics.

[0083] Using data-driven PID control algorithm, you don't need to rely on the precise mathematical model of the controlled object. You can calculate the parameters of the current PID control algorithm by using only the historical and current set values ​​and real-time values ​​as well as the historical PID control parameters. It has a strong ability to adapt to system changes.

[0084] S140, adjusting the speed of the air compressor or the opening degree of the solenoid valve to increase or decrease the air supply volume to achieve the purpose of constant pressure air supply.

[0085] Figure 2This is a gas load curve of a workshop in one day provided by an embodiment of the present invention. There are 3 variable frequency screw compressors in the workshop, 2 in use and 1 in standby. The maximum gas output of each air compressor is 40m 3 / min; Figure 3 This is a schematic diagram of the PID control effect based on data drive.

[0086] Figure 3 The effect curve diagram of the method of the present invention applied to a certain workshop, wherein (a) is the relationship curve between the pressure of the gas storage tank and time, and (b) is the relationship curve between the frequency of the variable frequency screw air compressor and time. It can be seen from the figure that the set pressure of each screw machine is 0.665MPa. The control method of the present invention can ensure that the pressure of the gas storage tank fluctuates between 0.66 and 0.67MPa, while in an air compression station without a control system, the pressure of the gas storage tank fluctuates between 0.65 and 0.75MPa. Therefore, the present invention significantly reduces the fluctuation of the pressure band. At the same time, the frequency of each variable frequency screw air compressor unit changes in the high-efficiency range, ensuring the efficient operation of the air compression system under various gas loads.

[0087] The present invention combines data-driven technology with PID control strategy to more flexibly respond to pressure change requirements under various complex working conditions, thereby improving the stability and efficiency of the entire compressed air supply network. It can be widely used in various industrial and commercial scenarios to improve the level of automation and energy conservation and emission reduction effects.

Claims

1. A variable frequency screw air compressor system control method based on data driven PID control, characterized in that: The following steps are involved: Obtaining the pressure value and set pressure value of the gas storage tank, calculating the adjustment amount of the pressure value of the gas storage tank by using a data-driven PID control algorithm, and determining the speed of the air compressor or the opening degree of the solenoid valve according to the adjustment amount; The control parameters of the data-driven PID control algorithm are calculated using the difference between the pressure value of the gas storage tank and the set pressure value to reduce the difference between the pressure value of the gas storage tank and the set pressure value.

2. A variable frequency screw air compressor system control method based on data driven PID control according to claim 1, characterized in that: The data-driven PID control algorithm adopts the PI control algorithm structure.

3. The control method of a variable frequency screw air compressor system based on data driven PID control according to claim 2 is characterized in that: The PI control algorithm structure is: Δu(t)=K I (t)e(t)-K P (t)Δy(t), Among them, u(t) is the adjustment amount of the gas tank pressure value at the current moment, K I (t) and K P (t) is the control parameter of the PI control algorithm, K I (t) is the ratio of the PI control algorithm at the current moment, K P (t) is the integral gain of the PI control algorithm at the current moment, y(t) is the pressure value of the gas tank at the current moment, e(t)=r(t)-y(t) is the difference between the set pressure value at the current moment and the pressure value of the gas tank, r(t) is the set pressure value at the current moment, and t represents the current moment.

4. The control method of a variable frequency screw air compressor system based on data driven PID control according to claim 3 is characterized in that: The control parameter calculation process of the PI control algorithm includes: The PI control parameters determined based on historical experience are used as the initial database; Calculate the distance between the information vector at the current moment and the information vector in the initial database; Select k groups of information vectors with the smallest distance from the initial database; Using k groups of information vectors, linear weighting is used to calculate the estimated value of the PI control parameter vector at the current moment; The steepest descent method is used to optimize the estimated value of the PI control parameter vector at the current moment to obtain a new PI control parameter vector; Update the new PI control parameter vector to the database.

5. The control method of a variable frequency screw air compressor system based on data driven PID control according to claim 4 is characterized in that: The expression of the initial database is: Φ(j)=[H(j),η(j)],j=1,2,…,N(0), Where N(0) is the number of information vectors stored in the initial database, H(j) = [r(j+1), r(j), y(j), …, y(jn y +1),u(j-1),…,u(jn u +1)] is an information vector, including r(j+1) is the set pressure value at time j+1, r(j) is the set pressure value at time j, y(j) is the gas tank pressure value at time j, u(j-1) is the adjustment amount of the gas tank pressure value at time j-1, and the information length of the information vector is n y +n u +1, and n y ≥3 and n u ≥2, n y is the output order, n u is the order of the input, η(j)=[K P (j),K I (j)] is the PI control parameter vector at time j, K P (j) is the integral gain of the time-varying controller at time j, K I (j) is the ratio of the time-varying controller at time j, and j represents time j.

6. A variable frequency screw air compressor system control method based on data driven PID control according to claim 5, characterized in that: Calculate the distance between the information vector at the current moment and the information vector in the initial database. The expression is: Where N(t) is the number of information vectors in the database at the current moment, H(t) is the information vector at the current moment, l represents the lth element, and H l (t) is the lth element of the information vector at the current moment, H l (j) is the lth element of the jth information vector, maxH l (m) is the maximum value of the lth element of all information vectors in the database, minH l (m) is the minimum value of the lth element of all information vectors in the database.

7. A variable frequency screw air compressor system control method based on data driven PID control according to claim 6, characterized in that: The expression of the estimated value of the PI control parameter vector at the current moment is: Among them, ω i represents the weight of the i-th information vector, and η(i) is the i-th PI control parameter vector.

8. The control method of variable frequency screw air compressor system based on data driven PID control according to claim 7 is characterized in that: The expression of the new PI control parameter vector is: Among them, λ is the learning factor, which is an adjustable parameter; is the performance index, y r (t+1) is the output at time t+1, To control the gradient of the performance index to the PI parameters.

9. A variable frequency screw air compressor system control method based on data driven PID control according to claim 8, characterized in that: The weight of the i-th information vector is expressed as:

10. An electronic device, characterized in that: include: at least one processor; and, a memory communicatively coupled to at least one of the processors; In which, the memory stores instructions that can be executed by at least one of the processors, and the instructions are executed by at least one of the processors so that at least one of the processors can execute a variable frequency screw air compressor system control method based on data-driven PID control as described in any one of claims 1 to 9.

Images