An adaptive shafting alignment control method for a complex airbag vibration isolation system

Abstract

The application discloses a kind of self-adapting shafting centering control methods for complex air bag vibration isolation system, initialize data set D is empty set, set the upper limit P of each air bag vibration isolator safe working pressure valueP l And lower limit P u , determine the control target Γ of air bag vibration isolator and optimal control target Γ opt ;Get the air bag pressure and control target information of N air bag vibration isolators at current time t, and add the collected data to data set D;Update correlation coefficient matrix R;Judge whether optimal control target Γ opt Has been reached under current system state, select the control target that needs to be controlled, according to the control target that needs to be controlled, carry out safety constraint check, and select the air bag vibration isolator number that needs to be controlled according to correlation coefficient;Without establishing complex system model, according to the correlation coefficient between the control object and control target generated by the data collected, then follow the safety constraint to adaptively complete system control, and it is suitable for variable working conditions and operating environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of industrial control, in particular to a self-adaptive shafting alignment control method for a complex air bag vibration isolation system. BACKGROUND

[0002] In recent years, air bag vibration isolation systems have been widely used in the field of ships due to their high-efficiency mechanical noise reduction and vibration isolation performance. An air bag vibration isolation system generally comprises a group of air bag isolators and a control system. The air bag isolators have low natural frequency and are highly adjustable, and can provide high-efficiency wide-frequency vibration isolation performance. The control system can monitor the state of the system and actively adjust the air bag pressure according to the control strategy to achieve the goal of adjusting the attitude of the bearing equipment. In particular, when the bearing equipment is a ship shafting equipment, the control system can assist in completing high-precision shafting alignment control by monitoring the alignment state, effectively solving the problem of difficult alignment in ship shafting vibration isolation.

[0003] However, with the increasing demand for vibration isolation, vibration isolation systems are developing towards integration. From single-machine vibration isolation to whole-cabin vibration isolation, the types of bearing objects are diversified, and the excitation sources are also complex, which puts higher requirements on the intelligence level of vibration isolation devices. In large and complex vibration isolation systems, there are factors such as raft deformation, complex changes in bearing equipment combination working conditions, and equipment aging, which make the alignment control of complex air bag vibration isolation systems a challenge.

[0004] On the one hand, in the existing air bag vibration isolation system alignment control method, it is generally assumed that the system is a linear time-invariant system. The control response mechanism model of the system is established according to the mass center position, weight and other information of the bearing equipment. However, the mass center position, weight and other information of the bearing equipment cannot be accurately obtained, and with factors such as equipment aging, the physical characteristics of the bearing equipment will change, so the mechanism model constructed will have certain deviation from the actual system, thereby causing inaccurate alignment control or non-convergent control. On the other hand, the existing control method is based on pre-set rules to carry out control, and when the operating conditions or environment change greatly, the control effect will also deteriorate due to the inability of the established rules to adapt to the dynamically changing environment. SUMMARY

[0005] In order to overcome the above technical defects, the present application provides a self-adaptive alignment control method for a complex air bag vibration isolation system, which has strong adaptability and good control effect.

[0006] The self-adaptive shafting alignment control method for a complex air bag vibration isolation system provided by the present application comprises N air bag isolators, and each air bag isolator is inflated and deflated by a control electromagnetic valve. The alignment control method is as follows:

[0007] S0) initialization, initialize the data set D as an empty set, set the upper limit P of the safe working pressure value of each air bag isolator l and the lower limit P u , determine the control target Γ of the air bag isolator and the optimal control target Γ opt ;

[0008] The control target Γ includes M control targets, denoted as:

[0009] Γ={τ1,…,τ M};

[0010] The optimal control target Γ opt is denoted as: Γ opt ={τ1 opt ,…,τ M opt};

[0011] The upper and lower limits of the pressure value are P l ={p1 l ,…,p N l} and P u ={p1 u ,…,p N u} respectively;

[0012] S1) Obtain the air bag pressure P(t)={p1,…,p N} and the control target information Γ(t)={τ1,…,τ M} of the N air bag isolators at the current time t, and add the collected data to the data set D;

[0013] S2) Update the correlation coefficient matrix R;

[0014] S3) Determine whether the optimal control target Γ opt has been reached under the current system state, if the optimal control target Γ opt has been reached, jump to step S1, otherwise jump to step S4;

[0015] S4) Select the control object: select the control target that needs to be controlled, perform safety constraint check according to the control target that needs to be controlled, and select the air bag isolator number that needs to be controlled according to the correlation coefficient;

[0016] S5) Perform the control action, and then jump to step S1.

[0017] Further, in the step S2), the correlation coefficient is denoted as:

[0018] R={r ij |1≤i≤N,1≤j≤M}

[0019] wherein r ij represents the correlation between the pressure of the ith air bag vibration isolator and the jth control target;

[0020] If the number of data entries K in the data set D={(P1,Γ1),…,(P K ,Γ K )} is less than the set threshold K T , the correlation coefficient is set to r ij =1;

[0021] Otherwise, when K≥K T , the function formula for calculating the correlation coefficient is:

[0022]

[0023] wherein, represents the pressure value of the ith air bag vibration isolator in the kth data in the data set D={(P1,Γ1),…,(P K ,Γ K )}; represents the average value of the pressure value of the ith air bag vibration isolator in the data set; represents the jth control target component in the kth data in the data set D; represents the average value of the jth control target in the data set.

[0024] Further, the condition for achieving the optimal control target in step S3) is:

[0025]

[0026] That is, each variable in the control target should satisfy the accuracy constraint, wherein ξ j is a constant.

[0027] Further, the specific process of step S4) is:

[0028] S41) Calculate the difference between the current control target and the optimal control target:

[0029]

[0030] S42) Select the control target with the largest absolute value of the deviation from the optimal control target, denoted as Δτ m =max{|Δτ j |},1≤j≤M, wherein m represents the sequence number in the set;

[0031] S43) Formulate a control strategy according to the safety constraints of the system:

[0032] If Δτ m < 0, the inflation action is preferred: first calculate the maximum control target bias Δτ m The corresponding auxiliary inflation decision value is denoted as The calculation method is as follows:

[0033]

[0034] If The elements in are not all zero, set the flag bit flag charge = 0, and calculate the auxiliary inflation decision probability distribution The calculation method is as follows:

[0035]

[0036] If The elements in are all zero, calculate the maximum control target bias Δτ m The corresponding auxiliary deflation decision value is denoted as The calculation method is as follows:

[0037]

[0038] If The elements in are not all zero, set the flag bit flag charge = 1, and calculate the auxiliary deflation decision probability distribution The calculation method is as follows:

[0039]

[0040] If The elements in are all zero, set the flag bit flag charge = 2;

[0041] If Δτ m > 0, the deflation action is preferred: first calculate the maximum control target bias Δτ m The corresponding auxiliary deflation decision value is denoted as The calculation method is as follows:

[0042]

[0043] If The elements in are not all zero, set the flag bit flag charge = 1, and calculate the auxiliary deflation decision probability distribution The calculation method is as follows:

[0044]

[0045] If If all the elements in the are zero, then calculate the maximum control target bias Δτ m The corresponding auxiliary inflation decision value is denoted as The calculation method is as follows:

[0046]

[0047] If If all the elements in the are not zero, set the flag bit flag charge = 0, and calculate the auxiliary inflation decision probability distribution The calculation method is as follows:

[0048]

[0049] If If all the elements in the are zero, set the flag bit flag charge = 2.

[0050] Further, in the step S5), if flag charge = 0, an inflation action is performed; according to the probability distribution value of , the number of the air bag vibration isolator that needs to be inflated is selected, assuming that the selected is the i-th air bag vibration isolator, then the air bag vibration isolator is inflated by Δp; Δp is a preset fixed value.

[0051] If flag charge = 1, a deflation action is performed; according to the probability distribution value of , the number of the air bag vibration isolator that needs to be deflated is selected, assuming that the selected is the i-th air bag vibration isolator, then the air bag vibration isolator is deflated by Δp, Δp is a preset fixed value.

[0052] If flag charge = 2, the control action cannot be performed, and an exception is reported.

[0053] The present application adopts the correlation coefficient to describe the correlation between the control object and the control target, avoids the problem of inaccurate complex model modeling, and through the collection of operation data, the correlation coefficient is updated in real time, and is fed back to the control algorithm for real-time adjustment, so that the best control strategy can be adaptively calculated.

[0054] The present application has the following technical effects:

[0055] 1) The air bag vibration isolation system adaptive control method provided by the present application does not need to establish a complex system model, generates the correlation coefficient between the control object and the control target according to the collected data, and then adaptively completes the system control according to the safety constraint, and is strong in adaptability to variable working conditions and operating environments.

[0056] 2) The correlation coefficient used by the application is updated in real time by the collected system operation data, without manual intervention, and has high intelligence;

[0057] 3) The control method provided by the application has simple structure, small calculation amount, is convenient to implement and deploy, and has good control effect. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 is a schematic diagram of the composition of a typical air bag vibration isolation system;

[0059] Figure 2 is a schematic diagram of the flow of the control method of the application;

[0060] Figure 3 is a training curve of the control effect of the application in a simulation environment. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical scheme and advantages of the application clearer and more apparent, the application will be further described in detail below in combination with the drawings and examples. It should be understood that the specific examples described herein are only used to explain the application and do not limit the application.

[0062] As shown in Figure 1 , a typical air bag vibration isolation system generally includes a set of air bag vibration isolators and a control system. The air bag vibration isolators can directly or indirectly support the load-carrying equipment through a raft support. The control system mainly includes an acquisition module, a calculation module and a control module. The acquisition module is responsible for acquiring the operating state of the system, the working condition of the equipment, etc., such as the centering state of the shafting equipment. The calculation module is responsible for logical operation and control strategy formulation. The control module is responsible for executing control actions, such as completing the inflation / deflation action of the air bag vibration isolator through the inflation / deflation control unit, achieving the attitude control of the raft, and then adjusting the centering state of the shafting equipment.

[0063] The adaptive shaft centering control method for a complex air bag vibration isolation system includes the following steps, as shown in Figure 3 :

[0064] S0) Initialization. The data set D is initialized as an empty set. Assuming that the air bag vibration isolation system includes 8 air bag vibration isolators, each air bag vibration isolator is inflated and deflated by a control solenoid valve. Then, the combination of air bag pressure values is represented as P=(p1,...,p8); for the safe working pressure of the air bag, a safety constraint is set, assuming that the upper and lower limits of the pressure values of all air bag vibration isolators are the same, which are 2.0Mpa and 0.8Mpa respectively, then the safety constraint can be represented as P l ={p1 l ,…,p8 l}={0.8Mpa,…,0.8Mpa}、P u={p1 u ...,p8 u} = {2.0 MPa, ..., 2.0 MPa}. Determine the control objective Γ and the optimal control objective Γ. opt Taking shaft alignment control as an example, the control objective generally consists of four components: vertical offset τ1, horizontal offset τ2, vertical skew τ3, and horizontal skew τ4, represented as: Γ={τ1,τ2,τ3,τ4}. The optimal control objective is that both the offset and skew values ​​are 0, i.e.: Γ opt ={0,0,0,0}.

[0065] S1) Obtain the airbag pressure P(t) = {p1,…,p8} and the corresponding control target information Γ(t) = {τ1,…,τ4} at the current time t of the airbag vibration isolation system; and add the collected data to the dataset D.

[0066] S2) Update the correlation coefficient matrix R. The system has a total of 8 airbag isolators and 4 alignment state components. Therefore, calculate the correlation between the airbag isolator pressure values ​​and the 4 alignment states. The correlation coefficient is expressed as:

[0067] R = {r ij |1≤i≤8,1≤j≤4}

[0068] Where, r ij This represents the correlation between the pressure of the i-th airbag and the j-th control target component.

[0069] If the dataset D = {(P1,Γ1),…,(P…} K ,Γ K The number of data entries K in the array is less than the set threshold K. T Then set the correlation coefficient to r. ij =1.

[0070] Otherwise, when K≥K T The function for calculating the correlation coefficient is:

[0071]

[0072] in, The dataset D represents the dataset D = {(P1,Γ1), ...,(P...}. K ,Γ K The pressure value of the i-th airbag isolator in the k-th data in )} This represents the average pressure value of the i-th airbag isolator in the dataset. This represents the j-th control target component in the k-th data point of dataset D. Let represent the average value of the j-th control objective component in the dataset. In a typical airbag vibration isolation system prototype, the updated correlation matrix is:

[0073]

[0074] where, r 11 =-0.12 indicates that there is a negative correlation between the pressure of the No. 1 air bag isolator and the vertical offset, and the correlation value is 0.12.

[0075] S3) Determine whether the optimal control target Γ has been reached under the current system state opt , if the optimal control target Γ has been reached opt , jump to step S1, otherwise jump to step S4;

[0076] The condition for reaching the optimal control target is:

[0077]

[0078] That is, each variable in the control target should satisfy the accuracy constraint, where ξ j is a constant. In engineering applications, it is generally considered that the system is in a centered state when the vertical / horizontal offset is less than 0.5 mm and the vertical / horizontal deflection is less than 0.5 mm / m. Therefore, ξ j = 0.5.

[0079] S4) Select the control object. Select the control target component that needs to be controlled. According to the target component that needs to be controlled, perform safety constraint checking, and select the air bag isolator number that needs to be controlled according to the correlation coefficient.

[0080] S41) Calculate the difference between the current control target and the optimal control target:

[0081]

[0082] Assume that the difference between the current centered state and the optimal centered state is: ΔΓ = {2.04, -2.40, 0.12, 3.71}.

[0083] S42) Select the component with the largest absolute value of the control target and the optimal control target deviation, denoted as Δτ m = max{|Δτ j |},≤1j≤M, where m represents the serial number in the set. At this time, the largest absolute value of the deviation is the fourth component, that is, the horizontal deflection Δτ4 = 3.71, m = 4.

[0084] S43) Develop a control strategy according to the safety constraints of the system. At this time, if Δτ m = 3.71 > 0, prefer to perform the deflation action. First, calculate the auxiliary deflation decision value corresponding to the maximum target deviation Δτ m , denoted as The calculation method is as follows:

[0085]

[0086] Based on the maximum offset and the correlation matrix, and given the current airbag isolator pressure value as P = {1.5, 1.5, 1.5, 1.5, 1.5, 1.5, 1.5, 1.5}, the following can be calculated:

[0087]

[0088] because If not all elements in the array are zero, then set the flag. charge =1, and calculate the probability distribution of auxiliary venting decisions. The calculation method is as follows:

[0089]

[0090] The probability distribution of the auxiliary deflation decision is obtained as follows:

[0091]

[0092] S5) Execute the control action, and then jump to step S1.

[0093] If flag charge =0, inflate the air. According to... The probability distribution value is used to select the number of the airbag isolator that needs to be inflated. Assuming that the selected airbag isolator is number i, then the airbag isolator is inflated by Δp, where Δp is a preset fixed value.

[0094] If flag charge =1, execute the deflation action. According to The probability distribution value is used to select the number of the airbag isolator that needs to be deflated. Assuming that the selected airbag isolator is number i, then the airbag isolator is deflated by Δp, where Δp is a preset fixed value.

[0095] If flag charge =2, unable to execute control action, exception reported.

[0096] Considering the flag obtained in step S4 charge =1, here we are based on The probability distribution is used to select the airbag isolator number that needs to be deflated. The probability of airbag isolator number 1 being selected is 31%, airbag isolator number 2 is selected is 26%, and so on. Assuming airbag isolator number 2 is selected, then airbag isolator number 2 will be deflated, reducing its air pressure by Δp = 0.1 MPa.

[0097] Figure 3 The shown implementation results are the control process of the centering control of the air bag isolation system using the adaptive control algorithm. Assuming that the control is performed once per second, it can be seen that, at about 60 seconds, the controller can quickly achieve centering control from the shaft system misalignment state, reach the control target, and the control effect is relatively stable.

Claims

1. A method for adaptive shafting alignment control for complex air-bag isolation system, the air-bag isolation system comprises N air-bag isolators, each air-bag isolator is inflated and deflated by controlling solenoid valve; characterized in that: The centering control method is specifically as follows: S0) initialization, initialize the data set D as an empty set, set the upper limit P of the safe working pressure value of each air bag isolator l and the lower limit P u , determine the control target Γ of the air bag isolator and the optimal control target Γ opt ; The control target Γ includes M control targets, denoted as: Γ = {τ1, ···, τ M}; Optimal control objective Γ opt is denoted as: Γ opt = {τ1 opt , ···, τ M opt} ; The upper and lower limits of the pressure value are P l = {p1 l , ···, p N l} and P u = {p1 u , ···, p N u , respectively. S1) Obtain the airbag pressure P(t) of N airbag vibration isolators at the current time t = {p1, ..., p N } and control target information Γ(t)={τ1,···,τ M }; and add the collected data to dataset D; S2) Update the correlation coefficient matrix R, and the correlation of the correlation coefficient matrix R represents the correlation between the pressure of the air bag vibration isolator and the control target; S3) determining whether the optimal control goal G has been reached in the current system state opt If yes, jump to step S1, otherwise jump to step S4 opt ​ S4) Select the control object: select the control target that needs to be controlled, perform safety constraint checking according to the control target that needs to be controlled, and select the air bag vibration isolator number that needs to be controlled according to the correlation coefficient; S5) Perform a control action, and then jump to step S1.

2. The adaptive shafting centering control method for a complex air bag vibration isolation system according to claim 1, wherein in step S2), the correlation coefficient is denoted as: R = {r ij |1≤i≤N,1≤j≤M} wherein r ij represents the correlation between the pressure of the i-th air bag vibration isolator and the j-th control target; If the number of data entries K in the data set D = {(P1, Γ1), · · ·, (P K , Γ K )} is less than a set threshold K T , then set the correlation coefficient as r ij = 1. Otherwise, when K≥K T The function for calculating the correlation coefficient is: wherein represents the i-th air spring isolator pressure value in the k-th data in the data set D = {(P1, Γ1), · · ·, (P K , Γ K )}, represents the average of the i-th air spring isolator pressure value in the data set; represents the j-th control target component in the k-th data in the data set D, represents the average of the j-th control target in the data set.

3. In the adaptive shaft alignment control method for complex airbag vibration isolation systems according to claim 2, in step S3), the optimal control objective is achieved. The condition is expressed as: i.e. each variable in the control target should satisfy the accuracy constraint, where ξ j is a constant.

4. The adaptive shafting centering control method for a complex air bag vibration isolation system according to claim 3, wherein the specific process of step S4) is: S41) Calculate the difference between the current control target and the optimal control target: S42) Select the control target with the largest absolute value of the deviation of the control target from the optimal control target, denoted as Δτ m = max{|Δτ j |}, 1≤j≤M, where, m represents the sequence number in the set; S43) Formulate a control strategy according to the safety constraints of the system: If Δτ m < 0, the inflation action is performed preferentially: first calculate the maximum control target offset Δτ m The corresponding auxiliary inflation decision value is denoted as The calculation method is as follows: If the elements in are not all zero, set the flag bit flag charge = 0 and compute the auxiliary aeration decision probability distribution The computation method is as follows: If all elements in the vector are zero, calculate the maximum control target bias Δτ m The corresponding auxiliary bleed decision value, denoted as The calculation method is as follows: If the elements in the vector are not all zero, set the flag bit flag charge = 1 and compute the auxiliary bleed decision probability distribution The computation method is as follows: If all the elements in are zero, set the flag bit flag charge = 2; If Δτ m > 0, the bleed action is performed preferentially: first calculate the maximum control target offset Δτ m The corresponding auxiliary bleed decision value is denoted as The calculation method is as follows: If not all elements in are zero, set flag charge = 1 and compute the auxiliary bleed decision probability distribution The computation method is as follows: If all elements in are zero, then calculate the maximum control target bias Δτ m The corresponding auxiliary inflation decision value, denoted as The calculation method is as follows: If the elements in are not all zero, set the flag bit flag charge = 0 and compute the auxiliary aeration decision probability distribution The computation method is as follows: If all elements in are zero, set flag charge = 2.

5. The adaptive shafting alignment control method for a complex air bag vibration isolation system according to claim 4, wherein in step S5), if flag charge = 0, an inflation action is performed; and a number of air bag vibration isolators that need to be inflated is selected according to the probability distribution value, and assuming that the selected number is i, the i-th air bag vibration isolator is inflated by Δp; Δp is a preset fixed value. ​ If flag charge = 1, execute the deflation action; according to the probability distribution value, select the air bag damper number that needs to be deflated, and assume that the selected is the i number air bag damper, deflate the air bag damper Δp, and Δp is a preset fixed value; If flag charge = 2, control action cannot be performed, report exception.

Images