CN111673026B - Online control method and control system for pressing process of forging press - Google Patents
Online control method and control system for pressing process of forging press Download PDFInfo
- Publication number
- CN111673026B CN111673026B CN202010388228.8A CN202010388228A CN111673026B CN 111673026 B CN111673026 B CN 111673026B CN 202010388228 A CN202010388228 A CN 202010388228A CN 111673026 B CN111673026 B CN 111673026B
- Authority
- CN
- China
- Prior art keywords
- forging press
- control
- value
- pressing process
- state information
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B21—MECHANICAL METAL-WORKING WITHOUT ESSENTIALLY REMOVING MATERIAL; PUNCHING METAL
- B21J—FORGING; HAMMERING; PRESSING METAL; RIVETING; FORGE FURNACES
- B21J9/00—Forging presses
- B21J9/10—Drives for forging presses
- B21J9/20—Control devices specially adapted to forging presses not restricted to one of the preceding subgroups
Abstract
The invention discloses an online control method and system for a pressing process of a forging press. The control method comprises the following steps: acquiring state information of sampling points at a first moment, randomly selecting opening degree control quantity of an adjusting mechanism, acquiring state information of the sampling points at a second moment, calculating the difference between the absolute values of the actual speed and the set speed at the two moments, and acquiring an instantaneous reward value; estimating the overall value of the pressing process to obtain the time sequence errors of two moments; deriving the relation between the control quantity and the state variable in the pressing process, and obtaining the optimal value of the opening degree of the regulating mechanism in the current state by using a tabu algorithm so as to further cause the piston to change; and returning to the step 3 to repeatedly execute convergence until the value of the opening control quantity of the adjusting mechanism is not changed any more. The invention realizes the online control of the forging press by combining reinforcement learning and tabu search, reduces the dependence on the model accuracy and simultaneously improves the working efficiency.
Description
Technical Field
The invention belongs to the field of online control of forging presses, and particularly relates to an online control method and an online control system suitable for a pressing process of a high-precision forging press.
Background
The existing hydraulic forging press can realize the quality control of the forged piece by adopting high-precision control methods such as sliding mode control, adaptive control and the like, thereby reducing the difficulty of subsequent machining process. However, the control effect of the method is greatly influenced by the accuracy of the model, and the model is inaccurate along with the extension of the service life of the model, which leads to the control accuracy not meeting the functional requirement, and how to find a control method capable of automatically matching with the environmental change is an urgent problem to be solved.
Although the existing data acquisition system and intelligent sensor technology can conveniently acquire the data of the forging press in real time, the high-precision forging press is mainly applied to some special forging occasions, even some experimental places, and enough effective data are difficult to obtain as a training set, so that the online application mode is difficult to popularize in the high-precision forging press based on offline learning.
Different from an off-line training and on-line control mode based on a neural network and a support vector machine, reinforcement learning is used as a third learning mode, and is a method for simulating animals to adapt to external environment and self-adjust. The learning mode can adopt two modes of off-line training and on-line training, the time sequence difference is taken as the correction deviation, and the stage reward is approximated through the adjustment of the behavior, so that the optimal control behavior based on the current data is finally obtained. From the perspective of engineering application, the reinforcement learning control method is applied to robots, unmanned automobiles and the like.
The main problems faced by reinforcement learning are the expression of the mapping relationship between the state space and the action space and the slow learning algorithm. Learning methods of reinforcement learning are classified into learning based on evaluation and learning based on policy iteration, and currently, methods based on policy iteration are mainly classified into three categories: (1) and (3) policy iteration method: firstly, estimating a value function under the current policy, and then carrying out policy improvement; (2) policy gradient method: estimating the gradient of the prediction return from the sample strategy by adopting an estimator; (3) the method for optimizing without derivative comprises the following steps: the return is used as a black box model, and optimization is directly carried out according to policy parameters, and the common methods comprise the following steps: cross-entropy (CEM) and Covariance Matrix Adaptation (CMA).
Because of the unknown external environment, the online learning of reinforcement learning usually adopts trial and error, and requires enough trial and error times to obtain the mapping relationship. Too long training time is often intolerable for real-time systems, and because the control effect cannot be predicted in advance during training, the danger of system runaway exists.
Tabu Search (TS for short) is a global gradual optimization algorithm, which avoids circuitous Search by introducing a flexible storage structure and corresponding Tabu criteria, and prevens some Tabu superior states by scofflaw criteria, thereby ensuring diversified efficient exploration to finally achieve global optimization. Meanwhile, the development of computer simulation and simulation technology provides a method which can predict the state of the system under the condition of not implementing actions, thereby avoiding the danger of runaway caused by the control actions applied to the system.
Therefore, an online control method combining data and a model, which is suitable for a high-precision forging press, is needed to be provided, so that the problem that the stability of a system cannot be controlled in the learning process of the traditional method is solved, the learning efficiency of online reinforcement learning is improved, and the calculation amount is reduced.
Disclosure of Invention
The invention aims to overcome the defects in the prior art and provides an online control method for the pressing process of a forging press. The method utilizes reinforcement learning to establish a reward mechanism, improves the reinforcement learning speed to a certain extent through a tabu search method, controls the forging hydraulic press to train and work on line simultaneously based on a data driving mode, reduces the dependence on model accuracy, and improves the work efficiency.
The online control method for the pressing process of the forging press is an online control method aiming at the combination of data and a model of the pressing process of the forging press on the basis of the existing forging press model. The semi-solid metal forming process using the forging press comprises six processes of rapid pressing, slow pressing, pressure maintaining, slow lifting and rapid lifting, and the "pressing process" of the forging press referred to herein refers to a process from the contact of a slide block of the forging press to a forging to be processed until the start of a pressure maintaining stage, i.e., a pressing process.
An on-line control method for a pressing process of a forging press comprises the following steps:
step 1: updating the state information of the sampling points at intervals of preset time on a preset forging press model, and acquiring the state information of the sampling points at the first moment by using a sensing system in the pressing process and sending the state information to a Trio controller of a control system; the sampling point state information comprises the working slide block pressure, the working slide block displacement and the working slide block speed in the current state of the forging press pressing process;
step 2: randomly selecting an opening control quantity of an adjusting mechanism, transmitting the opening control quantity to the Trio controller, and controlling an opening value of the adjusting mechanism through the Trio controller;
and step 3: acquiring state information of the sampling point at the second moment by using a sensing system;
and 4, step 4: calculating the difference between the absolute values of the actual speed and the set speed at the first time and the second time to obtain an instant reward value;
and 5: estimating the whole value (reward) of the pressing process on the basis of the forging press model, namely obtaining a deviation value from the pressing start to the pressing end; calculating the time sequence error of the sampling point state information at the first moment and the second moment;
step 6: and (3) deriving the relation between the control quantity and the state variable of the pressing process by utilizing the Ri Asia Ponuo stability theorem according to the obtained time series error, wherein the following conditions are met:
then, a taboo algorithm is used for obtaining an optimal value of the opening degree of the adjusting mechanism in the current state, and the opening degree control quantity of the adjusting mechanism is adjusted according to the obtained optimal value, so that the piston is further changed; otherwise, continuously trial-calculating until the conditions are met;
wherein u represents a proportional servo valve opening degree control amount, K represents a Young's modulus of an equal volume of oil, B represents a viscous damping coefficient, and KnRepresenting the nominal flow gain, PsIndicating the constant delivery pump output pressure, ΔpnIndicating the pressure loss of the valve port, ωnIndicating the natural frequency, F, of the proportional servo valvelRepresenting resistance to load, X2Indicating proportional servo valve input pressure, x5Indicating the upper chamber pressure, x, of the cylinder4Indicating proportional servo valve flow;
and 7: and returning to the step 3, repeatedly executing the steps 3-7 for convergence until the value of the opening control quantity of the adjusting mechanism is not changed any more, and ending.
Further, the tabu algorithm includes:
performing field search according to the current state quantity parameters, generating a field optimization solution, judging by using a scofflaw criterion, if the scofflaw criterion is met, using the optimal value of the opening degree of the adjusting mechanism in the current state as a historical optimal value, taking the historical optimal value out of a taboo table, keeping the historical optimal value unchanged, and putting the optimal value of the opening degree of the adjusting mechanism in the current state into a preset taboo table;
when a termination condition is reached (usually, a preset number of search steps is reached or a preset time is reached), a historical optimum value, that is, an optimum value of the opening degree control amount of the adjustment mechanism in the current state is obtained.
Wherein the adjusting mechanism is a proportional servo valve in the forging press.
An online control system for a pressing process of a forging press comprises: the forging press comprises a control system, a sensing system, an adjusting mechanism and a forging press.
The control system comprises an industrial personal computer, a display, a communication bus, a Trio controller and a PLC module, wherein the industrial personal computer is used for communicating with the Trio controller and the PLC module through the communication bus; the Trio controller is used for receiving sampling point state information transmitted by the sensing system, performing control algorithm calculation according to the received state information and controlling the opening value of a proportional servo valve of the forging press according to a calculation result, wherein the control algorithm comprises a reinforcement learning algorithm and a taboo algorithm, and the PLC module is used for realizing the electric logic control of the forging press; and the Trio controller and the PLC module realize data sharing through bus technology communication and transmit the data to the industrial personal computer through the communication bus.
The sensing system comprises a pressure sensor, a flow sensor and a displacement sensor, wherein the pressure sensor is used for detecting the upper cavity pressure of a hydraulic cylinder of the forging press and transmitting the upper cavity pressure to the Trio controller, the displacement sensor is used for transmitting the displacement of a working slide block of the slide block mechanism in the forging press to the Trio controller, and the flow sensor is used for transmitting the pipeline flow of the forging press to the Trio controller.
The adjusting mechanism is a proportional servo valve of the forging press, and the opening value of the proportional servo valve is adjusted according to the opening control quantity calculated by the Trio controller so as to adjust the flow of the valve and further complete the control of the speed of the working slide block.
The forging press, comprising: the oil source mechanism, the actuating mechanism, the control mechanism and the safety and auxiliary mechanism;
the forging press comprises an oil source mechanism, an oil source mechanism and a forging press mechanism, wherein the oil source mechanism comprises an oil source and an oil pump, a motor in the forging press drives the oil pump to press oil in the oil source into an oil pipe, and meanwhile, the hydraulic oil is pressurized, and the hydraulic oil with the obtained working pressure transmits force to an execution mechanism through a pipeline so as to press a forging piece;
the actuating mechanism comprises a hydraulic cylinder and a sliding block mechanism, the hydraulic cylinder drives a working sliding block in the sliding block mechanism to directly act on the forge piece, and the forge piece is pressed at different pressing speeds according to process requirements.
The control mechanism comprises a switch valve group and a proportional servo valve, the switch valve group is used for realizing the logic function of the forging process, and the proportional servo valve controls the valve opening value of the proportional servo valve through the Trio controller so as to adjust the valve flow and further complete the control of the speed of the working slide block.
The safety and auxiliary mechanism is used for ensuring the safety of the whole system and realizing other auxiliary work except pressing, such as top die, working surface movement and the like.
Compared with the prior art, the invention has the following beneficial effects:
1. the online control method is an online control method combining data and a model, the traditional offline learning is avoided, the learning and the control can be carried out together by utilizing the online control mode of the learned model, and the online optimization control of the forging press is realized;
2. the problem that the stability of the system cannot be controlled in the learning process by the traditional method is solved by establishing a system stability condition as a constraint condition through a model, so that the system stability of the system in the whole process is ensured;
3. the method for introducing tabu search improves the elimination of the redundant state irrelevant to the current state in the state space, thereby improving the efficiency of reinforcement learning.
Drawings
FIG. 1 is a schematic structural diagram of an on-line control system for the pressing process of a forging press according to the present invention;
FIG. 2a is a schematic diagram of the pressing process speed of a forging press in a constant speed state by using the on-line control method of the present invention;
FIG. 2b is a schematic of the press process controller output at a constant speed;
FIG. 3a is a schematic drawing of the pressing process speed at variable speed for a forging press employing the on-line control method of the present invention;
FIG. 3b is a graph of the output of the press process controller in a variable speed condition;
FIG. 4a is a schematic diagram of the pressing process speed of the forging press adopting the online control method of the invention at different sampling periods; FIG. 4b is a schematic diagram of the output of the forging press in different sampling periods according to the online control method of the invention.
Detailed Description
The technical solutions of the present invention are further described in detail with reference to the drawings and specific embodiments, which are only illustrative and not intended to limit the present invention.
The general approach to reinforcement learning is to use an iterative approach until convergence. This approach results in a slow training process due to the large number of iterative processes required. In fact, once the state information x (k) is detected, the controllable quantity is limited in a certain feasible domain due to the limitation of the characteristics of the system, the feasible domain can be directly searched without performing a large amount of iterative operations, and the feasible domain is unknown and lacks of a clear representation mode, so that the feasible domain can be searched only by adopting a random search method.
An online control method for a pressing process of a forging press comprises the following steps:
step 1: updating the state information of the sampling points at intervals of preset time on a preset forging press model, wherein the preset time is 2-5 min, acquiring the state information x (k) of the sampling points at the first moment by using a sensing system in the pressing process and sending the state information x (k) to a Trio controller of a control system, wherein the state information of the sampling points comprises the pressure of a working slide block, the displacement of the working slide block, the speed of the working slide block and the like in the current state of the pressing process of the forging press;
step 2: randomly selecting an action u (k), namely randomly selecting the opening control quantity of a proportional servo valve in a control mechanism of the forging press, wherein the opening control quantity ranges from 0V to 10V, transmitting the opening control quantity to a Trio controller, and controlling the opening value of the proportional servo valve through the Trio controller;
and step 3: acquiring sampling point state information x (k +1) at a second moment by using the sensing system;
and 4, step 4: calculating the difference between the absolute values of the actual speed and the set speed at the first time and the second time to obtain an instantaneous reward value, namely obtaining the pressing speed of the action:
the feasible range of the control quantity is [2 ] for n being AD converter limited by the number of bits of the hardware DA converter-n，2n]The press speed of a forging press is determined by the material properties, and generally requires a constant speed press over a range of temperatures, or a press speed that meets the requirements of certain process profiles. For this purpose, the real-time value is selected as the difference between the absolute values of the actual speed and the set speed at two times
R(k)＝||v(k)-vset(k)|-|v(k+1)-vset(k+1)|| (1)
And 5: estimating the whole value (reward) of the pressing process on the basis of the forging press model, namely obtaining a deviation value from the pressing start to the pressing end; and calculating the time sequence error of the state information of the sampling points at the first moment and the second moment.
According to equation (1), the cost function Vk(x, u) and Vk+1(x +1, u) is readily obtained from the coarse model
Note that the coarse model is more trending accurate than the TD (0) algorithm, so the target choice for tabu search is minc (x) minu (Vk (x, u) -Vk +1(x +1, u) + r (k)) (4).
Step 6: according to the obtained time sequence error, the relation between the control quantity and the state variable in the pressing process is derived by utilizing the Riemerov stable theorem, if the opening degree of an output valve meets the safe working stability condition of the forging press, the optimal value of the valve opening degree in the current state is obtained by using a tabu algorithm, and the control quantity of the valve opening degree is adjusted according to the obtained optimal value, so that the piston is further changed; otherwise, trial calculation is continuously carried out until the conditions are met.
Wherein the relationship between the state variable and the controlled variable can be derived from the Liya Ponux stability theorem
V＝xTPx (5)
Where P is a semi-positive array of the form
According to the physical meaning x of the state variableiNot equal to 0(i ═ 2,3,4, 5, 6), so
V＝xTPx＞0 (6)
If I is less than or equal to 0 and II is less than 0, the whole equation meets the stability of Liya Punuo.
For I and II
I＝ATP+PA≤0 (8)
II＝uT gT(x)Px+xT PgT(x)＜0 (9)
Bringing formula (5) into (9) to obtain
Solve (10) to obtain (11)
Get
Bringing formula (12) into (9) to obtain
Solve (13) to obtain (14)
Wherein u represents a valve opening control amount, K represents the Young's modulus of the oil with equal volume, B represents a viscous damping coefficient, and KnRepresenting the nominal flow gain, PsIndicating the constant delivery pump output pressure, Δ pnIndicating the pressure loss of the valve port, ωnIndicating the natural frequency, F, of the proportional servo valvelRepresenting load resistance, X2 representing proportional servo valve input pressure, X5 representing cylinder upper chamber pressure, X4 representing proportional servo valve flow, and R representing a coefficient
ρ represents the oil density, l represents the pipe length, S1 represents the pipe cross-sectional area, K represents the Young' S modulus of an equal volume of oil, S2 represents the piston cross-sectional area of the cylinder output chamber, B represents the viscous damping coefficient, K represents the viscous damping coefficientnIndicating the nominal flow gain, ξ the damping rate of a proportional servo valve, m the mass of the slider, λcRepresenting the leakage coefficient of the cylinder, V0The volume of the upper cavity of the hydraulic cylinder.
The obtained formula (14) constitutes a constraint that the control quantity, the state quantity and the load guarantee the stability of the system, and the constraint implies the relationship between the quantities, thereby narrowing the scope for tabu search.
The embodiment adopts a basic tabu search method to explain the application of the method. For an element X in a discrete space X, our goal is
min C(x)
s.t.x∈X (15)
Neighborhood s (x) is the set of solutions reachable by domain moves, which are shown in equation (17)
s(x)＝x+wd (16)
Where w is the step size and d is the direction. The tabu table is used for preventing the search from circulating, recording a plurality of previous walking points, directions or target values, and forbidding return. The tabu table is dynamically updated in a first-in first-out manner. Craving level function A (s, x) is a value dependent on s and x, and if there is C (s (x)) < A (s, x), then s (x) is not restricted by the T table. Even if s (x) e T, x ═ s (x) can still be taken.
The taboo algorithm in this embodiment has the following steps:
step 1: giving an algorithm parameter, randomly generating an initial solution x, and setting a tabu table to be null;
step 2: is it determined whether the algorithm termination condition is satisfied? If yes, finishing the algorithm and outputting an optimization result; otherwise, continuing;
and step 3: generating all (or a plurality) neighborhood solutions by utilizing the neighborhood function of the current solution work, and determining a plurality of candidate solutions from the neighborhood solutions;
and 4, step 4: is the scofflaw criterion satisfied for the candidate solution? If yes, replacing x with the optimal state y meeting the scofflaw criterion to become a new current solution, namely x is equal to y, replacing the taboo object which enters the taboo table earliest with the taboo object corresponding to y, and replacing the state of 'best so far' with y, and then turning to the step 6; otherwise, the following steps are continued.
And 5: judging the taboo attribute of each object corresponding to the candidate solution, selecting the optimal state corresponding to the non-taboo object in the candidate solution set as a new current solution, and simultaneously replacing the taboo object element which enters the taboo table earliest by the taboo object corresponding to the current solution.
Step 6: and (6) turning to the step 2.
When a termination condition is reached (usually, a preset number of search steps is reached or a preset time is reached), a historical optimum value, that is, an optimum value of the proportional servo valve opening degree control amount in the current state is obtained.
And 7, returning to the step 3, repeatedly executing convergence until the value of the opening control quantity of the proportional servo valve does not change any more, and ending.
FIG. 1 shows an online control system for a pressing process of a forging press, which comprises: the forging press comprises a control system, a sensing system, an adjusting mechanism and a forging press.
The control system comprises an industrial personal computer (not shown in the figure), a display, a communication bus, a Trio controller and a PLC module, wherein the industrial personal computer is used as an upper computer system of the control system and displays a monitoring picture of the working state of the forging press through the display, and the industrial personal computer is used for communicating with the Trio controller and the PLC module through the communication bus; the Trio controller is used for receiving sampling point state information transmitted by the sensing system, performing control algorithm calculation according to the received state information and controlling the opening value of a proportional servo valve of the forging press according to a calculation result, wherein the control algorithm comprises a reinforcement learning algorithm and a taboo algorithm, and the PLC module is used for realizing the electric logic control of the forging press; and the Trio controller and the PLC module realize data sharing through bus technology communication and transmit the data to the industrial personal computer through the communication bus.
The control system adopts S7-300 PLC to realize electrical logic control, adopts the trio controller MC224 to realize control algorithm, and can adopt other servo controllers to realize control algorithm in actual use; the control algorithm comprises reinforcement learning and taboo algorithms; and the trio controller and the PLC realize data sharing through Modbus bus technology communication and transmit the data to an upper computer system through Profibus.
The sensing system comprises a pressure sensor, a flow sensor and a displacement sensor, wherein the pressure sensor is used for detecting the upper cavity pressure of a hydraulic cylinder of the forging press and transmitting the upper cavity pressure to the Trio controller, the displacement sensor is used for transmitting the displacement of a working slide block of the slide block mechanism in the forging press to the Trio controller, and the flow sensor is used for transmitting the pipeline flow of the forging press to the Trio controller.
The adjusting mechanism is a proportional servo valve of the forging press, and the opening value of the proportional servo valve is adjusted according to the opening control quantity calculated by the Trio controller so as to adjust the flow of the valve and further complete the control of the speed of the working slide block.
The forging press, comprising: the oil source mechanism, the actuating mechanism, the control mechanism and the safety and auxiliary mechanism.
The forging press comprises an oil source mechanism, an oil source mechanism and a forging press mechanism, wherein the oil source mechanism comprises an oil source and an oil pump, a motor in the forging press drives the oil pump to press oil in the oil source into an oil pipe, the hydraulic oil is pressurized, and the hydraulic oil with the obtained working pressure transmits force to an execution mechanism through a pipeline so as to press a forging;
the actuating mechanism comprises a hydraulic cylinder and a sliding block mechanism, the hydraulic cylinder drives a working sliding block in the sliding block mechanism to directly act on the forge piece, and the forge piece is pressed at different pressing speeds according to process requirements. The pressing process of the invention is particularly from the contact of the working slide block with the forging until the pressing is finished.
The control mechanism comprises a switch valve group and a proportional servo valve, the switch valve group is used for realizing the logic function of the forging process, the proportional servo valve controls the valve opening value of the proportional servo valve through the Trio controller so as to adjust the valve flow, and further the control on the speed of the working slide block is completed, the proportional servo valve adopts a Leishile product, and the response time is less than 10 ms.
The safety and auxiliary mechanism is used for ensuring the safety of the whole system and realizing other auxiliary work except pressing, such as top die, working surface movement and the like.
The method and system of the present invention was tested using an ultra low speed forging hydraulic press as a test bed, as shown in fig. 2 through 4. The parameters of the online control system model are shown in table 1:
table 1: parameter value table of test system
Example 1: low constant speed state
The slide block pressing speed is set to be 0.03mm/s, the flow of oil required by the upper cavity of the hydraulic cylinder is small due to the slow pressing speed, and the opening of the proportional servo valve generates large reducing loss pressure due to the small opening, so that the opening of the proportional servo valve is a compromise between the working pressure of the hydraulic cylinder and the pressure loss of the hydraulic cylinder flowing through the servo valve. Since the tabu search as a random search may show a certain randomness, the present embodiment tests the pressing process 5 times, fig. 2a is a speed diagram, fig. 2b is an output diagram of the controller, and different colors represent each test.
It can be seen from fig. 2b that the curves are not completely covered, which reflects that the control values are not completely consistent each time by the method of the present invention, and fig. 2a reflects the effect of the control values on the index slider pressing speed, it can be seen from the figure that the deviations of the curves are not large, the maximum peak value is 0.0302, the minimum peak value is 0.0298 in the first test, and the relative error is 0.7% in the fourth test, which indicates that the different control values can meet the control requirement.
It is worth pointing out that the conventional PID control can achieve such a speed control accuracy, but the parameter adjustment of PID is very difficult, and the method proposed by us can automatically obtain the control value according to the detected state.
Example 2: state of speed change
The speed was set to 0.08mm/s from the beginning, 0.04mm/s through and 0.06mm/s at the end, and the process was described as shown in equation (17)
This example was repeated 5 times to test the effectiveness of the proposed method, and FIG. 3a is a graph of the press process speed in a variable speed condition; fig. 3b is a schematic diagram of the output of the controller.
It can be seen from fig. 3a that each curve tracks the set speed well during different speed changes, the speed set point is 0.08 during 1-30, the maximum peak value 0.0812 occurs at the 5 th, minimum value 0.0788; the speed setpoint was 0.04 during 50-80, with a maximum peak of 0.0406 occurring at time 3, a minimum of 0.0394, occurring at times 1,2,3, 4; the speed setpoint was 0.06, the maximum and minimum peaks were 0.0609 and 0.0394, respectively, with a relative error of 1.5% for the 80-100 periods. The mean and residual error of each run are shown in table 2.
Table 2: mean and residual table of speed and controller output in variable speed state
Example 3: influence of sampling period
This embodiment tests the effect of the recommended method on speed at different sample period lengths. The sampling period was chosen to be 1-5 minutes and the reference speed was chosen to be 0.04mm/s, the results obtained being shown in figures 4a and 4 b. The controller randomly selects action at the beginning and then enters automatic control according to the online control method. The curve in fig. 4a represents the slider depression speed at different sampling periods, and fig. 4b reflects the controller output at different sampling periods. As can be seen from the figure, different sampling periods have a large influence on the transient process, the sampling period is short, and the duration of the transient process is short; the sampling period is long and the transition time is longer. The main reason is that the control quantity obtained by the recommended method is not changed in a sampling period, when the sampling period is short, the method can be well adapted to the change of the system, and when the sampling period is long, the time for adapting to the change of the system is also prolonged. But the stable state can achieve better control effect, and the average value and the residual error in the stable state are shown in table 3.
Table 3: mean value and residual error table of speed and controller output under different sampling periods
| Sampling period (min) | Steady state process | Mean value | Residual error | |
| Speed 1 | 2 | 18min-90min | 0.0400 | 2.9115e-07 |
| Speed 2 | 3 | 25min-90min | 0.0401 | 5.6880e-07 |
| Speed 3 | 4 | 50min-90min | 0.0399 | 4.6431e-07 |
| Speed 4 | 5 | 70min-90min | 0.0399 | 3.9806e-07 |
While the present invention has been described with reference to the accompanying drawings, the present invention is not limited to the above-described embodiments, which are illustrative only and not restrictive, and various modifications which do not depart from the spirit of the present invention and which are intended to be covered by the claims of the present invention may be made by those skilled in the art.
Claims (4)
1. An online control method for a pressing process of a forging press is characterized by comprising the following steps:
step 1: updating the state information of the sampling points at intervals of preset time on a preset forging press model, acquiring the state information of the sampling points at a first moment by using a sensing system in the pressing process, and sending the state information to a Trio controller of a control system; the sampling point state information comprises the working slide block pressure, the working slide block displacement and the working slide block speed in the current state of the forging press pressing process;
step 2: the control method comprises the steps that the Trio controller obtains an opening control quantity of the adjusting mechanism and controls an opening value of the adjusting mechanism, wherein the opening control quantity is a randomly selected value and ranges from 0V to 10V;
and step 3: acquiring state information of a sampling point at a second moment by using the sensing system;
and 4, step 4: calculating the difference between the absolute values of the actual speed and the set speed at the first time and the second time to obtain an instant reward value;
and 5: estimating the overall value of the pressing process on the basis of the forging press model, namely obtaining a deviation value from the beginning to the end of pressing; calculating the time sequence error of the sampling point state information at the first moment and the second moment;
step 6: and (3) deriving the relation between the control quantity and the state variable of the pressing process by utilizing the Ri Asia Ponuo stability theorem according to the obtained time series error, wherein the following conditions are met:
wherein u represents a proportional servo valve opening degree control amount, K represents a Young's modulus of an equal volume of oil, B represents a viscous damping coefficient, and KnRepresenting the nominal flow gain, PsIndicating the constant delivery pump output pressure, Δ pnIndicating the pressure loss of the valve port, ωnIndicating the natural frequency, F, of the proportional servo valvelRepresenting resistance to load, X2Indicating proportional servo valve input pressure, x5Indicating the upper chamber pressure, x, of the cylinder4Indicating proportional servo valve flow;
then, obtaining an optimal value of the opening degree of the adjusting mechanism in the current state by using a tabu algorithm, and adjusting the control quantity of the opening degree of the adjusting mechanism according to the obtained optimal value, thereby further causing the piston of the forging press to change; otherwise, continuously trial-calculating until the conditions are met;
and 7: and returning to the step 3, repeatedly executing the steps 3-6 to converge until the value of the opening control quantity of the adjusting mechanism is not changed any more, and ending.
2. The on-line control method according to claim 1, wherein the tabu algorithm in step 6 comprises:
performing field search according to the current state quantity parameters, generating field optimization solutions, judging by using scofflaw criteria, if the scofflaw criteria are met, using the optimal value of the opening degree of the adjusting mechanism in the current state as a historical optimal value, taking the historical optimal value out of a taboo table, keeping the historical optimal value unchanged, and putting the optimal value of the opening degree of the adjusting mechanism in the current state into a preset taboo table;
when the termination condition is reached, a historical optimum value, that is, an optimum value of the opening degree control amount of the adjustment mechanism in the current state is obtained.
3. The on-line control method according to claim 1, wherein the adjusting mechanism is a proportional servo of a forging press.
4. An online control system for a pressing process of a forging press comprises: the forging press comprises a control system, a sensing system, an adjusting mechanism and a forging press; the method is characterized in that:
the control system comprises an industrial personal computer, a display, a communication bus, a Trio controller and a PLC module, wherein the industrial personal computer is used for communicating with the Trio controller and the PLC module through the communication bus; the Trio controller is used for receiving sampling point state information transmitted by the sensing system, performing control algorithm calculation according to the received state information and controlling the opening value of a proportional servo valve of the forging press according to a calculation result, wherein the control algorithm comprises a reinforcement learning algorithm and a taboo algorithm, and the PLC module is used for realizing the electric logic control of the forging press; the Trio controller and the PLC module are communicated through a bus technology to realize data sharing and are transmitted to the industrial personal computer through the communication bus;
the sensing system comprises a pressure sensor, a flow sensor and a displacement sensor, wherein the pressure sensor is used for detecting the upper cavity pressure of a hydraulic cylinder of the forging press and transmitting the upper cavity pressure to the Trio controller, the displacement sensor is used for transmitting the displacement of a working slide block of a slide block mechanism in the forging press to the Trio controller, and the flow sensor is used for transmitting the pipeline flow of the forging press to the Trio controller;
the adjusting mechanism is a proportional servo valve of the forging press, and adjusts the opening value of the proportional servo valve according to the opening control quantity calculated by the Trio controller so as to adjust the flow of the valve and further complete the control of the speed of the working slide block;
the forging press, comprising: the oil source mechanism, the actuating mechanism, the control mechanism and the safety and auxiliary mechanism;
the forging press comprises an oil source mechanism, an oil source mechanism and a forging press mechanism, wherein the oil source mechanism comprises an oil source and an oil pump, a motor in the forging press drives the oil pump to press oil in the oil source into an oil pipe, and meanwhile, the hydraulic oil is pressurized, and the hydraulic oil with the obtained working pressure transmits force to an execution mechanism through a pipeline so as to press a forging piece;
the actuating mechanism comprises a hydraulic cylinder and a sliding block mechanism, the hydraulic cylinder drives a working sliding block in the sliding block mechanism to directly act on the forge piece, and the forge piece is pressed at different pressing speeds according to process requirements;
the control mechanism comprises a switch valve group and a proportional servo valve, the switch valve group is used for realizing the logic function of the forging process, and the proportional servo valve controls the valve opening value through the Trio controller so as to adjust the valve flow and further complete the control of the speed of the working slide block;
the safety and auxiliary mechanism is used for ensuring the safety of the whole system and realizing other auxiliary work except pressing, and the other auxiliary work comprises the movement of the top die and the working surface.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010388228.8A CN111673026B (en) | 2020-05-09 | 2020-05-09 | Online control method and control system for pressing process of forging press |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010388228.8A CN111673026B (en) | 2020-05-09 | 2020-05-09 | Online control method and control system for pressing process of forging press |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111673026A CN111673026A (en) | 2020-09-18 |
| CN111673026B true CN111673026B (en) | 2022-04-01 |
Family
ID=72451908
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010388228.8A Active CN111673026B (en) | 2020-05-09 | 2020-05-09 | Online control method and control system for pressing process of forging press |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111673026B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113351817A (en) * | 2021-04-30 | 2021-09-07 | 天津市天锻压力机有限公司 | Flexible control method of forging press |
| CN116719358B (en) * | 2023-08-11 | 2023-10-10 | 梅州市博富能科技有限公司 | Cell pressing control system based on pressure detection |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3930160B2 (en) * | 1998-09-07 | 2007-06-13 | 本田技研工業株式会社 | Method and apparatus for controlling forging die |
| CN101819424B (en) * | 2010-01-25 | 2013-01-16 | 南京埃斯顿自动化股份有限公司 | Method and controller for controlling position, pressure measurement and safe control of sliding block of forging machine tool |
| CN103537599B (en) * | 2013-10-18 | 2015-11-18 | 中南大学 | For control method and the control system thereof of forging press |
| CN105880449B (en) * | 2016-04-07 | 2017-09-05 | 中南大学 | The on-line prediction method and system of drop press running status |
| CN105903874B (en) * | 2016-04-25 | 2018-02-09 | 太原科技大学 | A kind of high speed forging hydraulic press forging deformation work(compensating control method |
| KR20190048727A (en) * | 2017-10-31 | 2019-05-09 | 주식회사 엔제이시스템 | Control device for forging press |
-
2020
- 2020-05-09 CN CN202010388228.8A patent/CN111673026B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN111673026A (en) | 2020-09-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111673026B (en) | Online control method and control system for pressing process of forging press | |
| CN111176115B (en) | Valve position control method based on fuzzy neural network and humanoid intelligent control | |
| CN110703693B (en) | Iterative learning feedforward control method and system for machine tool feeding system | |
| CN110488759B (en) | Actor-Critic algorithm-based numerical control machine tool feed control compensation method | |
| CN109695533B (en) | Automatic intelligent setting method for neutral position of speed regulator hydraulic servo system | |
| CN103537599A (en) | Forging press control method and control system of forging press | |
| CN109352655B (en) | Robot deformation compensation method based on multi-output Gaussian process regression | |
| CN112659498B (en) | Real-time optimal control method for injection molding machine deep neural network | |
| CN108897228B (en) | Online control method and system for forging process | |
| Dai et al. | Adaptive force tracking control of electrohydraulic systems with low load using the modified LuGre friction model | |
| CN113934183B (en) | Electrohydraulic servo system friction compensation method based on improved firefly algorithm | |
| CN114548311B (en) | Hydraulic equipment intelligent control system based on artificial intelligence | |
| CN102873106B (en) | Quick and precise elongation control method for temper mill | |
| CN111796511A (en) | Wavelet neural network PID (proportion integration differentiation) online control method and system of hydraulic actuator | |
| CN115081763B (en) | Energy consumption prediction system for turning process of numerical control lathe | |
| CN114294277B (en) | Synchronization method of double-cylinder hydraulic hoist for opening and closing flood discharge gate based on oil pressure regulation | |
| Inoan | Movement control of an unloading machine from a rotary furnace | |
| CN107191154B (en) | Wellhead back pressure regulating method and device | |
| CN113110305A (en) | Friction modeling method of electromechanical system and application thereof | |
| Zhang et al. | An online control approach for forging machine using reinforcement learning and taboo search | |
| CN110685310A (en) | Engine extreme load control method and system based on neural network and engineering vehicle | |
| CN112947093B (en) | Distributed robust tracking control method, system and storage medium for robot | |
| Chen et al. | Master Cylinder Oil Pressure Following Control Algorithm of Electric Power Assisted Braking System Based on Fuzzy PID Controller | |
| CN117728735B (en) | Motor drive control method, device, equipment and storage medium | |
| CN117872717A (en) | Fuzzy self-adaptive setting PID control method for valve positioner |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |