CN116894465A - Injection molding machine charging barrel temperature prediction method - Google Patents
Injection molding machine charging barrel temperature prediction method Download PDFInfo
- Publication number
- CN116894465A CN116894465A CN202310325982.0A CN202310325982A CN116894465A CN 116894465 A CN116894465 A CN 116894465A CN 202310325982 A CN202310325982 A CN 202310325982A CN 116894465 A CN116894465 A CN 116894465A
- Authority
- CN
- China
- Prior art keywords
- neural network
- pso
- injection molding
- temperature
- molding machine
- 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.)
- Pending
Links
- 238000001746 injection moulding Methods 0.000 title claims abstract description 50
- 238000000034 method Methods 0.000 title claims abstract description 27
- 238000013528 artificial neural network Methods 0.000 claims abstract description 54
- 239000002245 particle Substances 0.000 claims abstract description 41
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 26
- 230000003044 adaptive effect Effects 0.000 claims abstract description 12
- 238000003062 neural network model Methods 0.000 claims abstract description 9
- 230000006870 function Effects 0.000 claims description 26
- 238000002347 injection Methods 0.000 claims description 16
- 239000007924 injection Substances 0.000 claims description 16
- 210000002569 neuron Anatomy 0.000 claims description 11
- 239000000243 solution Substances 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000001133 acceleration Effects 0.000 claims description 3
- 230000005284 excitation Effects 0.000 claims description 3
- 230000004044 response Effects 0.000 abstract description 10
- 238000012549 training Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 5
- 238000010438 heat treatment Methods 0.000 description 5
- 238000006243 chemical reaction Methods 0.000 description 4
- 230000001276 controlling effect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 3
- 238000013527 convolutional neural network Methods 0.000 description 3
- 238000013178 mathematical model Methods 0.000 description 3
- 239000011159 matrix material Substances 0.000 description 3
- 238000005457 optimization Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 229910000831 Steel Inorganic materials 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 239000000853 adhesive Substances 0.000 description 1
- 230000001070 adhesive effect Effects 0.000 description 1
- 230000003321 amplification Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 230000000875 corresponding effect Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000004132 cross linking Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 238000005485 electric heating Methods 0.000 description 1
- 238000005187 foaming Methods 0.000 description 1
- 239000008187 granular material Substances 0.000 description 1
- 239000002932 luster Substances 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000002844 melting Methods 0.000 description 1
- 230000008018 melting Effects 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 230000010355 oscillation Effects 0.000 description 1
- 238000003672 processing method Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000010845 search algorithm Methods 0.000 description 1
- 238000010008 shearing Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0499—Feedforward networks
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B29—WORKING OF PLASTICS; WORKING OF SUBSTANCES IN A PLASTIC STATE IN GENERAL
- B29C—SHAPING OR JOINING OF PLASTICS; SHAPING OF MATERIAL IN A PLASTIC STATE, NOT OTHERWISE PROVIDED FOR; AFTER-TREATMENT OF THE SHAPED PRODUCTS, e.g. REPAIRING
- B29C45/00—Injection moulding, i.e. forcing the required volume of moulding material through a nozzle into a closed mould; Apparatus therefor
- B29C45/17—Component parts, details or accessories; Auxiliary operations
- B29C45/76—Measuring, controlling or regulating
- B29C45/78—Measuring, controlling or regulating of temperature
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/048—Activation functions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/082—Learning methods modifying the architecture, e.g. adding, deleting or silencing nodes or connections
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Physiology (AREA)
- Manufacturing & Machinery (AREA)
- Mechanical Engineering (AREA)
- Injection Moulding Of Plastics Or The Like (AREA)
Abstract
The invention discloses a method for predicting the temperature of a charging barrel of an injection molding machine, which comprises the following steps: s1: initializing a topological structure of the BP neural network; s2: initializing a weight and a threshold of the BP neural network, and encoding the initialized weight and threshold into particles; s3: setting basic parameters of a PSO algorithm, and optimizing a BP neural network by using the PSO algorithm to obtain a PSO-BP neural network model; s4: introducing adaptive variation into a PSO-BP neural network model, and constructing and obtaining a PSO-BP neural network with the introduced adaptive variation; s5: predicting the temperature of a charging barrel of the injection molding machine by adopting the obtained PSO-BP neural network with the self-adaptive variation introduced; the invention can effectively achieve the aim of unifying the overshoot of the control temperature and the quick response by utilizing the obtained PSO-BP-PID control algorithm.
Description
Technical Field
The invention relates to the technical field of injection molding, in particular to a method for predicting the temperature of a charging barrel of an injection molding machine.
Background
The injection molding machine is a key device for processing plastic products, and the temperature is a very key control index in the processing process of the injection molding machine. In the injection molding process, the injection molding machine charging barrel is the only place for melting injection molding pieces, and is also a link with very strict requirements on precision in the transmission and electric heating parts of the whole injection molding machine.
The temperature of a charging barrel of the injection molding machine is too low, shearing force is generated in plastic among screws, cold curing is generated to damage the machine, and the problems of poor elasticity, luster and adhesive force of products and the like are caused; the temperature of the charging barrel is too high, the crosslinking reaction can occur among plastic molecules, so that the tissue is loose, the foaming phenomenon occurs, and meanwhile, the decomposition of materials can be accelerated. Proportional Integral Derivative (PID) control is simple and easy to implement, and can be widely applied to various temperature control systems. However, in terms of overshoot of control temperature and quick response, the conventional PID control method cannot unify the overshoot and the quick response.
The invention relates to a method for controlling the temperature of a charging barrel of an injection molding machine, which is applied to a method for intelligently controlling the operation scene of the injection molding machine by using a convolutional neural network to replace a PID controller of the charging barrel of the injection molding machine, as in Chinese patent CN110202768A, the publication date of 2019, and the date of 05 and 17; the main controlled variables in the operation process of the injection molding machine comprise parameters such as temperature, speed, pressure and the like, and most of the injection molding machines in the market control the injection molding process through a traditional PID controller; the method comprises the steps of converting a time sequence obtained by sampling an input/output signal in the operation control process of a charging barrel temperature PID controller of an injection molding machine into a matrix form, wherein the converted input/output matrix is expressed into a gray level diagram form, and elements of the matrix correspond to elements of a gray level image; and uploading the gray images to CNN in a centralized server for supervised training, wherein the trained CNN has better control effect compared with a charging barrel temperature PID controller of the injection molding machine. The temperature control method of the injection molding machine charging barrel used by the invention cannot achieve uniformity in the overshoot and quick response of the control temperature.
Disclosure of Invention
The invention aims to solve the technical problems that: the temperature control method of the injection molding machine charging barrel cannot achieve the unified technical problem in controlling the overshoot of the temperature and quick response. The invention provides a method for predicting the temperature of an injection molding machine charging barrel, which can achieve unification in controlling the overshoot of the temperature and quick response.
In order to solve the technical problems, the invention adopts the following technical scheme: a method for predicting the temperature of a charging barrel of an injection molding machine comprises the following steps:
s1: initializing a topological structure of the BP neural network;
s2: initializing a weight and a threshold of the BP neural network, and encoding the initialized weight and threshold into particles;
s3: setting basic parameters of a PSO algorithm, and optimizing a BP neural network by using the PSO algorithm to obtain a PSO-BP neural network model; s4: introducing adaptive variation into a PSO-BP neural network model, and constructing and obtaining a PSO-BP neural network with the introduced adaptive variation;
s5: and predicting the temperature of the injection molding machine charging barrel by adopting the obtained PSO-BP neural network with the self-adaptive variation introduced.
Firstly, initializing a topological structure of a BP neural network, determining the number of neurons of an input layer, an hidden layer and an output layer, and selecting an activation function of the neural network; selecting factors influencing the temperature of a charging barrel of the injection molding machine as input of the input layer, and taking an integral coefficient, a proportional coefficient and a differential coefficient as output of the output layer; and initializing a weight and a threshold of the BP neural network, encoding the initialized weight and threshold into particles, setting basic parameters of a PSO algorithm, optimizing the BP neural network by using the PSO algorithm, introducing adaptive variation into a PSO-BP neural network model, constructing the PSO-BP neural network introducing the adaptive variation, and adopting the PSO-BP neural network of the adaptive variation to predict convexity of strip steel and realize on-line prediction of the temperature of a charging barrel of an injection molding machine on an injection molding site.
Preferably, the step S4 includes the steps of:
s41: initializing the position and the speed of the particles according to a calculation formula;
s42: calculating an initial fitness value of the particles;
s43: updating the speed and the position, and recalculating the fitness value after the self-adaptive algorithm;
s44: and judging whether the iteration stop condition is met.
Particle and velocity initialization: after initializing the topological structure of the BP neural network, determining an initial weight and a threshold value, encoding the initial weight and the threshold value into particles, and setting basic parameters of a PSO algorithm;
calculating an initial fitness value of the particles: randomly inputting a particle, calculating an initial fitness value of the particle, comparing the obtained fitness value with an individual optimal value, and if the fitness value is larger than the individual optimal value, recording the current position of the particle, and taking the current position as the individual optimal position of the particle; then, continuously inputting all the remaining particles, and repeating until the algorithm is finished, so as to obtain the optimal position of the particle group;
updating the speed and the position, and recalculating the fitness value after the self-adaption algorithm: after updating the speed and the position of the particles, judging whether the speed and the position of the particles are in an allowable range, and updating the speed and the position of the particles according to the updated speed and the updated position; re-calculating the fitness value of the particles according to the self-adaptive variation, and updating the individual optimal value and the group optimal value of the particles according to the new fitness value of the particles;
judging whether an iteration stop condition is met: after iteration is finished, judging whether the obtained group optimal value meets the requirements of the weight and the threshold value of the BP neural network, and if so, finishing; if the requirement is not met, the weight and the threshold of the network are re-initialized, the fitness value of the particles is calculated again, and the algorithm is iterated until the condition is met.
Preferably, the BP neural network in the step S1 uses a PID structure, the BP neural network includes three layers, and an excitation function expression of an output layer of the BP neural network is as follows:
y＝g(x)＝1/(1+e -x )
the output layer neuron input is x, and the output layer neuron output is y. From this equation, specific data of the output layer can be calculated.
Preferably, the calculation formula of the step S41 is as follows:
v i (t+1)＝wv i (t)+c 1 rand()(P best，i -P i (t))+c 2 rand()(P best，i -P i )
P i (t+1)＝P i (t)+v i (t+1)
wherein: v i (t) updating the ith particle dimensionless speed t times; p (P) i (t) is a dimensionless location; c 1 ，c 2 Is an acceleration coefficient; random number with random () of 0-1; w is the inertia coefficient; p (P) best，i An individual optimal solution is found for the individual particles. Substituting the initial data into the equation to obtain the initial position and speed of the particle.
Preferably, the input layer of the BP neural network includes three neurons, and the output layer of the BP neural network includes three neurons. The BP neural network using the PID structure comprises three layers, namely an input layer, an hidden layer and an output layer.
Preferably, the input layer is the main influencing factor of injection molding machine barrel temperature, and the main influencing factors comprise screw rotation speed, injection temperature and injection pressure. The screw rotation speed, the injection temperature and the injection pressure are used as three inputs of the BP neural network, and corresponding outputs can be obtained.
Preferably, the expression of the PSO-BP neural network introducing adaptive variation is as follows:
θ 1 (x)＝[K(x，x 1 )，…，K(x，x N )]β
wherein: θ 1 Indicating the temperature of the cartridge; beta represents the input of mixed kernel function extreme learning machineAnd outputting a weight vector. Substituting the finally obtained output value and the output weight vector into the formula to obtain the temperature value of the charging barrel.
Preferably, the fitness value of the particle is judged by a fitness function, wherein the fitness function is as follows:
wherein: o (o) pi Hiding the output of the node for the ith sample, x pi Is the input of the ith sample. The quality of the obtained fitness is judged by the formula.
The invention has the following substantial effects: the invention optimizes the upgraded BP neural network mode by using the PSO algorithm to adjust each parameter of PID control, and obtains the PSO-BP-PID control algorithm, which can effectively achieve the aim of unifying the overshoot of control temperature and quick response.
Drawings
FIG. 1 is a flow chart of an embodiment;
FIG. 2 is a flowchart of step S4 of the embodiment;
FIG. 3 is a block diagram of a BP neural network of an embodiment;
FIG. 4 is a simulation diagram of the PSO algorithm process of an embodiment;
FIG. 5 is a simulation diagram of setting PID parameters using BP neural network and PSO-BP neural network according to an embodiment;
FIG. 6 is a PID control step response graph of a BP neural network and a PSO-BP neural network according to an embodiment;
FIG. 7 is a table comparing predicted values and actual values of cartridge temperatures for different predictive models of an embodiment.
Detailed Description
The following description of the embodiments of the present invention will be made with reference to the accompanying drawings.
In the method for predicting the temperature of the cylinder of the injection molding machine, as shown in fig. 1 and 2, there are usually 3 to 6 heating sections on the cylinder of the injection molding machine, and in the heating position, the power supply is not controlled by a single switch, but by a solid state relay of each section. In general, a reaction curve method is used for determining a mathematical model of a charging barrel heating system, after the adjacent temperature coupling phenomenon of a heating section is ignored, a step response temperature reaction curve is used as a model basis, and then analysis and research are carried out through the shape of the reaction curve, so that a charging barrel temperature change mathematical model of each section of the injection molding machine is obtained, and the charging barrel temperature change mathematical model can be indirectly regarded as a first-order inertia environment plus pure hysteresis link, as follows:
wherein: t is an inertial time constant; k is an amplification factor; s is a Laplace transform factor; l is the lag time in s; g(s) is the transfer function of the heating system.
And selecting main influencing factors of the temperature of the injection molding machine charging barrel, such as screw rotation speed, injection temperature and injection pressure, as the input of a model, and constructing an injection molding machine charging barrel temperature prediction model of the mixed kernel function extreme learning machine based on a PSO algorithm.
The model input is P (n), P (n-1), P (n-2), θ amb (n)，θ amb (n-1)，θ amb (N-2), v (N), v (N-1), v (N-2), (n=1, 2, …, N). P (n) represents the injection pressure, P (n-1) represents the injection pressure at the last moment, and P (n-2) represents the injection pressure at the last two moments; θ amb (n) represents the injection temperature, θ amb (n-1) represents the injection temperature at the previous time, θ amb (n-2) represents the injection temperature at the last two moments; v (n) represents the screw speed, v (n-1) represents the screw speed at the last moment, and v (n-2) represents the screw speed at the last two moments. Normalizing training sample data (such as screw rotation speed, injection pressure and injection temperature) to obtain the following formula:
wherein: q is training data before normalization of the sample data;q' is training data normalized by the sample data; q min Minimum value for each training sample data; q max Maximum value for each training sample data.
Compared with other random search algorithms, the PSO algorithm has the characteristics of good robustness, high convergence speed, simplicity, easiness in implementation and the like, and can quickly find out the global optimal solution of the problem. A number of spatial particles are randomly generated, and the initial position and velocity of each particle is calculated according to the following equation:
v i (t+1)＝wv i (t)+c 1 rand()(P best，i -P i (t))+c 2 rand()(P best，i -P i )
P i (t+1)＝P i (t)+v i (t+1)
wherein: v i (t) updating the ith particle dimensionless speed t times; p (P) i (t) is a dimensionless location; c 1 ，c 2 Is an acceleration coefficient; random number with random () of 0-1; w is the inertia coefficient; p (P) best，i An individual optimal solution is found for the individual particles.
The expression of the PSO-BP neural network introducing adaptive variation is as follows:
θ 1 (x)＝[K(x，x 1 )，…，K(x，x N )]β
wherein: θ 1 Indicating the temperature of the cartridge; beta represents the mixed kernel function extreme learning machine output weight vector.
In general, the conventional nonlinear regression method does not consider differences in negative and positive error processing methods for predicting the cartridge temperature. Negative errors can result in a less than complete and objective estimate of the barrel temperature, making the injection molding machine thermal state estimation scheme inadequate for demand. Meanwhile, the positive error can realize safe, reliable and conservative thermal state of the injection molding machine, which is beneficial to safe operation of the injection molding machine. Therefore, the correction error is correlated by using larger tolerance, so that the prediction error is always in a positive value, and the prediction result is scientific and reliable. And combining the Gaussian kernel function and the polynomial kernel function to generate a novel mixed kernel function, so that the performance of the kernel function extreme learning machine is improved. Granules and method for producing the sameThe composition of the sub-position vector (u) includes the blending kernel parameter sigma 1 ，σ 2 Lambda, and mixed kernel function extreme learning machine output weight vector, see formula
u[β，σ 1 ，σ 2 ，λ]
According to the actual implementation mode of the particle swarm, continuously carrying out updating operation on the particle swarm, searching an optimal feasible solution (u) of the minimum fitness function, obtaining the output weight and the kernel function parameter of the kernel function extreme learning machine according to the determined u, determining the mapping network of the kernel function extreme learning machine, and completing the training purpose. The objective function is the mean square error (σ) of the predicted value, calculated as:
wherein: n is the number of predicted samples; θ 1 E represents the prediction result, and the constraint condition is theta 1，e ≥θ 1 The constraint condition can ensure that the predicted value of the temperature of the charging barrel of the injection molding machine is larger than the actual measured value to the greatest extent.
As shown in fig. 3, the BP neural network using the PID structure includes three layers, i.e., an input layer, an hidden layer, and an output layer. Wherein K is i Is an integral coefficient; k (K) p Is a proportionality coefficient; k (K) d Is a differential coefficient; x is x 1 ～x 3 Respectively set values of the 1 st section to the 3 rd section of the temperature of the charging barrel of the injection molding machine.
The control algorithm for PID is as follows:
u(l)＝u(l-1)+K p [e(l)-e(l-1)]+K i e(1)+K d [e(l)-2e(l-1)+e(l-2)]
wherein: 1 is the collection times; u (l) is the output; u (l-1) is output at the last moment; e (l) is the deviation; e (l-1) is the previous time deviation; e (l-2) is the deviation of the last two moments.
Input layer input signal O in BP neural network j (1) ＝x(j)，j＝1，2，3。
Due to PID structural algorithm and K in BP neural network i ，K p ，K d Positive numbers are guaranteed, and therefore,the excitation function of the BP neural network output layer is expressed as:
y＝g(x)＝1/(1+e -x )
the output layer neuron input is x, and the output layer neuron output is y.
PSO algorithm optimization can be divided into the following steps: (1) The BP neural network structure is determined by determining the number of nodes of an output layer and an input layer. The population size is 30, omega is 0.8, c1 and c2 are 2, and the maximum number of iterations is 50. (2) Calculating the fitness of the particles, and judging the position of the particles by establishing a fitness function, wherein in the BP neural network, an error function can be used as the fitness function, which is thatWherein: o (o) pi Hiding the output of the node for the ith sample, x pi Is the input of the ith sample. (3) calculating an individual optimum value and a global optimum value. And (4) updating the particle speed and the particle position. And (5) calculating the input and output of the BP neural network. (6) calculating the output of the PID controller. And (5) until the output value exceeds the maximum iteration number, obtaining an optimal solution.
As shown in fig. 4, an approximate mathematical calculation model is selected, and after 50 iterations, the adaptability change of the optimal individual of the PSO algorithm is shown in fig. 4. As can be seen from fig. 4: along with the increase of the iteration times, the adaptability of the PSO algorithm gradually tends to be stable, and after 50 iterations, the adaptability is basically unchanged.
As shown in fig. 5: compared with the BP neural network, the PSO-BP neural network of the PID adapts to the temperature control requirement of the charging barrel of the injection molding machine through training and learning, and the system has good dynamic performance and decoupling control performance.
As shown in fig. 6, PID parameters are determined by a BP neural network model and tuning is performed so that the variation amplitude of the PID parameters substantially coincides with the variation condition of the input signal law, but the tuning effect cannot meet the requirement because of significant oscillation fluctuation at the early stage. After PSO-BP neural network algorithm optimization, the fluctuation is obviously reduced, the performance of the PID controller is better, the fluctuation of the temperature of the charging barrel of the injection molding machine can be reduced to a higher degree, and meanwhile, the response speed and the accuracy of the temperature of the charging barrel are improved, so that the temperature of the charging barrel of the injection molding machine can be accurately controlled.
As shown in fig. 7, the effect of the conventional PID model on the temperature prediction of the cylinder is not ideal as a whole, and the fluctuation of the prediction result is large; the PSO-BP-PID neural network model has better overall prediction effect than the traditional PID model, and the overall trend of the temperature is similar to the actual value, so that the model has better fitting degree.
The foregoing examples illustrate only a few embodiments of the invention, which are described in detail and are not to be construed as limiting the scope of the invention. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the invention, which are all within the scope of the invention.
Claims (8)
1. The method for predicting the temperature of the charging barrel of the injection molding machine is characterized by comprising the following steps of:
s1: initializing a topological structure of the BP neural network;
s2: initializing a weight and a threshold of the BP neural network, and encoding the initialized weight and threshold into particles;
s3: setting basic parameters of a PSO algorithm, and optimizing a BP neural network by using the PSO algorithm to obtain a PSO-BP neural network model;
s4: introducing adaptive variation into a PSO-BP neural network model, and constructing and obtaining a PSO-BP neural network with the introduced adaptive variation;
s5: and predicting the temperature of the injection molding machine charging barrel by adopting the obtained PSO-BP neural network with the self-adaptive variation introduced.
2. The method according to claim 1, wherein the step S4 comprises the steps of:
s41: initializing the position and the speed of the particles according to a calculation formula;
s42: calculating an initial fitness value of the particles;
s43: updating the speed and the position, and recalculating the fitness value after the self-adaptive algorithm;
s44: and judging whether the iteration stop condition is met.
3. The injection molding machine barrel temperature prediction method according to claim 1, wherein the BP neural network in the step S1 uses a PID structure, the BP neural network includes three layers, and an excitation function expression of an output layer of the BP neural network is as follows:
y＝g(x)＝1/(1+e -x )
the output layer neuron input is x, and the output layer neuron output is y.
4. The method according to claim 2, wherein the calculation formula of step S41 is as follows:
v i (t+1)＝wv i (t)+c 1 rand()(P best,i -P i (t))+c 2 rand()(P best,i -P i )
P i (t+1)＝P i (t)+v i (t+1)
wherein: v i (t) updating the ith particle dimensionless speed t times; p (P) i (t) is a dimensionless location; c 1 ，c 2 Is an acceleration coefficient; random number with random () of 0-1; w is the inertia coefficient; p (P) best,i An individual optimal solution is found for the individual particles.
5. The injection molding machine barrel temperature prediction method of claim 1, wherein the input layer of the BP neural network comprises three neurons and the output layer of the BP neural network comprises three neurons.
6. The method of claim 5, wherein the input layer is a primary influencing factor for injection molding machine barrel temperature, the primary influencing factor comprising screw speed, injection temperature, and injection pressure.
7. A method for predicting the temperature of a cylinder of an injection molding machine according to claim 1,2 or 3, wherein the expression of the PSO-BP neural network introducing the adaptive variation is as follows:
θ 1 (x)＝[K(x,x 1 ),…,K(x,x N )]β
wherein: θ 1 Indicating the temperature of the cartridge; beta represents the mixed kernel function extreme learning machine output weight vector.
8. The method for predicting the temperature of a cylinder of an injection molding machine according to claim 2 or 4, wherein the fitness value of the particles is judged by a fitness function, and the fitness function is as follows:
wherein: o (o) pi Hiding the output of the node for the ith sample, x pi Is the input of the ith sample.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310325982.0A CN116894465A (en) | 2023-03-30 | 2023-03-30 | Injection molding machine charging barrel temperature prediction method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310325982.0A CN116894465A (en) | 2023-03-30 | 2023-03-30 | Injection molding machine charging barrel temperature prediction method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116894465A true CN116894465A (en) | 2023-10-17 |
Family
ID=88311256
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310325982.0A Pending CN116894465A (en) | 2023-03-30 | 2023-03-30 | Injection molding machine charging barrel temperature prediction method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116894465A (en) |
-
2023
- 2023-03-30 CN CN202310325982.0A patent/CN116894465A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107615186A (en) | The method and apparatus of Model Predictive Control | |
| CN111474965A (en) | Fuzzy neural network-based method for predicting and controlling water level of series water delivery channel | |
| CN111158237B (en) | Industrial furnace temperature multi-step prediction control method based on neural network | |
| CN112631215B (en) | Intelligent forecasting method, device, equipment and storage medium for industrial process operation index | |
| Kusumoputro et al. | System identification and control of pressure process rig system using Backpropagation Neural Networks | |
| CN110097929A (en) | A kind of blast furnace molten iron silicon content on-line prediction method | |
| CN108959787B (en) | Thermal deformation prediction method and system of macro-macro dual-drive system considering actual working conditions | |
| CN116306248A (en) | Lithium battery temperature field prediction method based on space-time nonlinear error compensation model | |
| CN113325696B (en) | Single neuron PID and model prediction combined hybrid control method applied to crosslinked cable production equipment | |
| CN111930010A (en) | LSTM network-based general MFA controller design method | |
| CN112379589A (en) | Worm wheel machining shape controllable process | |
| Zheng et al. | Adaptive fuzzy optimal control for switched nonlinear systems with output hysteresis | |
| CN111240201B (en) | Disturbance suppression control method | |
| CN116880191A (en) | Intelligent control method of process industrial production system based on time sequence prediction | |
| CN116894465A (en) | Injection molding machine charging barrel temperature prediction method | |
| CN110794676A (en) | CSTR process nonlinear control method based on Hammerstein-Wiener model | |
| CN114442557B (en) | Quick identification method and system for machine tool temperature field | |
| CN114880930A (en) | Entity equipment digital twin modeling simulation prediction method based on online learning | |
| CN112379601A (en) | MFA control system design method based on industrial process | |
| CN114330119A (en) | Deep learning-based pumped storage unit adjusting system identification method | |
| CN113625557A (en) | HVAC system model prediction control method of online optimization model | |
| CN113534661A (en) | Resistance furnace temperature control method based on Kalman filtering and non-minimum state space | |
| CN116149262B (en) | Tracking control method and system of servo system | |
| CN111781821B (en) | Parameter self-tuning method of SISO (SISO) compact-format model-free controller based on Attention mechanism cyclic neural network | |
| Yordanova et al. | Design and stability analysis of supervisor-based adaptive fuzzy logic control system for temperature |
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 |