PE plastic reaction prediction control method and system based on cyber-physical fusion

Abstract

The present invention discloses a predictive control method and system for PE plastic reactions based on cyber-physical fusion. The method comprises: establishing a deviation system model of the PE plastic reaction physical entity based on a target steady-state point based on knowledge of the PE plastic polymerization and separation mechanism based on nonlinear dynamics; summarizing the evolution of input signals in the deviation system based on communication security and physical device characteristics, and constructing a cyber-physical fusion prediction model for the PE plastic reaction deviation in combination with the deviation system model; designing performance indicators and control laws based on the cyber-physical fusion prediction model and prior knowledge of deception attacks, and establishing a minimization control optimization problem and constraints; calculating the minimization control optimization problem based on real-time discrete state information, and obtaining control instructions for regulating the physical entity in combination with the target production steady-state heat. The present invention can free the traditional PE plastic production industry from the limitations of the actual environment and hardware, providing more accurate, flexible, and rapid real-time perception and dynamic control for the production process, and ensuring the reliable and safe operation of PE plastic reactions and separations.

Description

Technical Field

[0001] The present invention belongs to the field of control technology, and in particular relates to a PE plastic reaction prediction control method, system, device, electronic device and computer-readable storage medium based on information-physics fusion. Background Art

[0002] Polyethylene (PE) plastic materials are widely used in various industries such as industry and power due to their outstanding advantages such as corrosion resistance, durability, and excellent electrical insulation. The PE production process is complex, and the slurry method is one of the main preparation methods. By adding a catalyst and raw materials to a reactor and thoroughly mixing and stirring, the ethylene monomer undergoes a polymerization reaction under the action of the catalyst. The resulting polymer is then subjected to flash separation, drying and granulation processes to finally obtain a high-density polyethylene product. The polymerization reaction is a highly exothermic process, and the temperature of the reactor has a significant impact on the quality and yield of the target PE product. Therefore, heat must be input or removed from the reactor in a timely manner. During the flash separation process, the separator needs to be provided with an appropriate temperature in combination with the boiling point of the target product to avoid the accidental distillation of non-target products, reduce impurities in the product, and improve product purity. Therefore, strict temperature control must be implemented for the polymerization reaction and flash separation process to keep the temperature in each reactor within the range required by the target process and improve the reaction and separation efficiency.

[0003] The traditional PE production industry is plagued by difficulties in management and control, and low levels of automation, resulting in high energy consumption and low economic returns. Production takes place under high-temperature and high-pressure conditions, and the materials involved are flammable and explosive, posing a risk of fire and explosion. Coupled with the complex on-site environment and limited local hardware capabilities, there is an urgent need to design effective intelligent control systems and temperature control methods to enhance production safety and automated control, ensuring reliable operation of the PE plastic chemical reaction process. Summary of the Invention

[0004] In order to address the deficiencies of the above-mentioned prior art, the present invention provides a PE plastic reaction predictive control method, system, device, electronic device and computer-readable storage medium based on information-physical fusion, which can enable the traditional PE plastic production industry to break free from the limitations of on-site environment and hardware, provide more accurate, flexible and rapid real-time perception and dynamic control of the reaction process, and ensure the reliable and safe operation of PE plastic reaction and separation.

[0005] The first object of the present invention is to provide a PE plastic reaction prediction and control method based on information-physics fusion.

[0006] The second object of the present invention is to provide a PE plastic reaction control system based on information-physics fusion.

[0007] The third object of the present invention is to provide a PE plastic reaction prediction and control device based on information-physics fusion.

[0008] A fourth object of the present invention is to provide an electronic device.

[0009] A fifth object of the present invention is to provide a computer-readable storage medium.

[0010] The first object of the present invention can be achieved by adopting the following technical solutions:

[0011] A PE plastic reaction prediction and control method based on cyber-physical fusion, the method comprising:

[0012] According to the knowledge of PE plastic polymerization and separation mechanism, a deviation system model based on the target steady-state point of the PE plastic reaction physical entity is established based on nonlinear dynamics;

[0013] Based on communication security and physical device characteristics, the evolution of input signals in the deviation system is summarized, and combined with the deviation system model, an information-physical fusion prediction model for PE plastic reaction deviation is constructed;

[0014] Based on the cyber-physical fusion prediction model and prior knowledge of deception attacks, performance indicators and control laws are designed, and the minimization control optimization problem and constraints are established;

[0015] The minimization control optimization problem is calculated based on real-time discrete state information, and the control instructions for regulating the physical entity are obtained in combination with the target production steady-state heat to achieve safe predictive control of the PE plastic reaction process.

[0016] Furthermore, the cyber-physical fusion prediction model and prior deception attack knowledge are used to design performance indicators and control laws, and to establish minimization control optimization problems and constraints, including:

[0017] Based on the cyber-physical fusion prediction model, a controlled output for controlling performance is constructed. is the state weight matrix, is the input weight matrix; x k is the state variable of the PE plastic reaction process deviation system at the kth sampling moment, u k is the input signal of the deviation system at the kth sampling moment;

[0018] The performance objective of solving the minimum maximization of the following expression for each sampling time k is:

[0019]

[0020] Where, the infinite time domain mean square performance index is represents the mathematical expectation, z k+i|kis the predicted value of the system controlled output at the kth sampling moment in the future k+i moment, and the control law adopts the linear state feedback control law u k+i|k =F k x k+i|k , F k is the state feedback control gain at the kth sampling moment, x k+i|k is the predicted value of the system state at the kth sampling moment in the future k+i moment, u k+i|k is the predicted value of the input signal output by the system controller at the k+ith moment in the future under the kth sampling moment;

[0021] The minimization control optimization problem and constraints are:

[0022] Given adjustable parameters ε1>0, ε2>0, 0<β<1, there exists a matrix Q k >0, G k 、Y k , Z k and the scalar γ k >0,δ k >0,φ k > 0 so that the following minimization problem for all j = 1, 2, ..., 2 nu Solvable:

[0023]

[0024] Security constraints:

[0025]

[0026] Q k ≥δ k I

[0027] Quadratic boundedness conditions:

[0028]

[0029] State constraints:

[0030]

[0031] Input saturation auxiliary constraints:

[0032]

[0033] Where, is the system matrix, is the input matrix; is the attack probability; Y k =F k G k , Z k =H kG k , X k,nn is the matrix X k The nth diagonal element of max,n is the absolute upper limit of the nth state, is the matrix Z k The rth row of , 0 is the zero matrix of the corresponding dimension, I is the identity matrix of the corresponding dimension, W is the Lipschitz matrix, E j is a diagonal matrix with diagonal elements of 0 or 1, n x is the dimension of the state variable, n u is the dimension of the input variables, is the maximum energy of a single attack; Obtain the feedback control gain.

[0034] Furthermore, the information-physics fusion prediction model of the PE plastic reaction deviation is:

[0035]

[0036] Where x k+1 、x k are the state variables of the cyber-physical fusion system of PE plastic reaction deviation at the k+1th and kth sampling moments respectively; is the system matrix; is the input matrix; α k is a random variable describing the link security; v k is the malicious false information injected by the attacker; sat(·) is the standard unit saturation function under scalar or vector; u k is the input signal at the kth sampling moment; g(·,·) is a nonlinear function.

[0037] Furthermore, based on the communication security and physical device characteristics, the evolution of the input signal in the deviation system is summarized, including the evolution of the input signal before and after communication and the evolution before and after execution by the actuator. The input signal actually executed by the actuator of the deviation system is obtained based on the two-step evolution.

[0038] The evolution of the input signal before and after communication is specifically:

[0039] For the deviation system, let u k is the input signal output by the controller, for u k After the input signal is transmitted by the communication unit, then:

[0040]

[0041] Where, αk =0 means the link is not attacked, α k =1 indicates that the link has been attacked; v k ∈Υ, n u is the dimension of the input vector, ||·|| 2 is the square of the L2 norm, v max is the maximum amplitude vector of a single attack, ζ is the upper bound of the total energy of the attack, so we have

[0042] The evolution of the executor before and after execution is as follows:

[0043] make Received by the actuator The input signal actually executed after is the execution input variable of the deviation system at the kth sampling moment, then:

[0044]

[0045] Furthermore, for the random attack process, the attack launch is modeled as a random sequence following the Bernoulli distribution. Based on historical attack records, the prior probability model is obtained:

[0046]

[0047] Where Prob{·} represents the probability, Determined based on past attack records.

[0048] Furthermore, based on the knowledge of the polymerization and separation mechanism of PE plastics, a deviation system model of the PE plastic reaction physical entity based on the target steady-state point is established based on nonlinear dynamics, including:

[0049] make is the state variable of the PE plastic reaction physical entity real system at the kth sampling moment, is the execution input variable of the PE plastic reaction physical entity real system at the kth sampling moment; is the target steady-state state of the physical entity, Input for the target steady-state execution of the physical body; As the state variable of the PE plastic reaction process deviation system at the kth sampling moment, is the execution input variable of the deviation system at the kth sampling moment;

[0050] The discrete nonlinear real system model of the PE plastic reaction process is linearized based on the target steady-state point, and the deviation system model of the PE plastic reaction process based on the target steady-state point is obtained as follows:

[0051]

[0052] Where x k+1 is the state variable of the PE plastic reaction process deviation system at the k+1th sampling time; is the linearized system matrix, is the linearized input matrix, It is the discrete nonlinear system function of the PE plastic reaction physical entity real system; It is the nonlinear function term of the deviation system of PE plastic reaction process.

[0053] Furthermore, the PE plastic reaction physical entity includes a flash separator and at least one polymerization reactor;

[0054] The discrete nonlinear real system model is obtained through the following process:

[0055] make is the state variable of the physical entity real system of the PE plastic reaction process, Input variables for the execution of physical reality systems;

[0056] The nonlinear dynamic model of the PE plastic reaction physical entity is discretized based on sampling time, and the discrete nonlinear real system model of the PE plastic reaction process is obtained as follows:

[0057]

[0058] Where, is the state variable at the k+1th sampling moment, is a discrete nonlinear function.

[0059] The second object of the present invention can be achieved by adopting the following technical solutions:

[0060] A PE plastic reaction control system based on cyber-physical fusion, the system includes a sensing unit, a control unit, an execution unit and a communication unit, wherein:

[0061] The sensing unit is used to collect data of the physical entity of the PE plastic industry reaction in real time and convert the collected data into discrete state information; the discrete state information is transmitted to the control unit through the communication unit;

[0062] A control unit is configured to store the received discrete state information and analyze it based on the above-mentioned PE plastic reaction predictive control method to obtain a control instruction for adjusting the physical entity; transmit the control instruction to the execution unit via the communication unit; the control instruction for adjusting the physical entity is the total heat required for the PE plastic reaction;

[0063] The execution unit is used to serialize the received control instructions and adjust the flow rate of the heat-carrying medium in the jacket by the actuator according to the serialized control instructions to achieve heating or cooling of the container, so that the polymerization and separation physical processes of PE plastic production can continue to operate normally;

[0064] The communication unit is used to realize communication between the control unit and the perception unit and execution unit.

[0065] Furthermore, the collected data includes the container temperature and the concentration of each material in the polymerization reaction and flash separation process.

[0066] The third object of the present invention can be achieved by adopting the following technical solutions:

[0067] A PE plastic reaction prediction and control device based on cyber-physical fusion, the device comprising:

[0068] The first building block is used to establish a deviation system model of the PE plastic reaction physical entity based on the target steady-state point based on nonlinear dynamics according to the knowledge of PE plastic polymerization and separation mechanism;

[0069] The second building block is used to summarize the evolution of input signals in the deviation system based on communication security and physical device characteristics, and to construct an information-physical fusion prediction model for PE plastic reaction deviation in combination with the deviation system model;

[0070] The third building block is used to design performance indicators and control laws based on the cyber-physical fusion prediction model and prior deception attack knowledge, and to establish the minimization control optimization problem and constraints;

[0071] The calculation module is used to calculate the minimization control optimization problem based on real-time discrete state information, and obtain the control instructions for regulating the physical entity in combination with the target production steady-state heat to achieve safe predictive control of the PE plastic reaction process.

[0072] The fourth object of the present invention can be achieved by adopting the following technical solutions:

[0073] An electronic device includes a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the above-mentioned PE plastic reaction prediction and control method based on information-physics fusion is implemented.

[0074] The fifth object of the present invention can be achieved by adopting the following technical solutions:

[0075] A computer-readable storage medium stores a program, which, when executed by a processor, implements the above-mentioned PE plastic reaction prediction and control method based on information-physics fusion.

[0076] The present invention has the following beneficial effects compared to the prior art:

[0077] 1. The method provided by the present invention realizes the dynamic control of the physical information system of the PE plastic reaction, ensuring the reliable and safe operation of the PE plastic reaction and separation. Taking into account the communication security issue, a priori probability model of deception attack is established according to historical attack situations; through the prediction model based on the information-physical fusion framework, the mean square significance performance target is designed to achieve predictive control under random attacks, enhance the resistance to communication attacks, and ensure the mean square safety of PE reaction temperature control. Taking into account the inherent physical limitations of the jacket control valve, a less conservative approach is used to solve the saturation nonlinearity of the actuator to improve the control performance; according to the real-time reaction state information, the predictive control problem is optimized in a rolling manner, the feedback controller gain is updated, and the accuracy of temperature control and the timeliness of adjustment are improved, which is of great significance for reducing material and energy consumption and improving the economic benefits of PE plastic production;

[0078] 2. This invention constructs a cyber-physical fusion control system for the PE plastics industrial reaction process. Relying on communication technology, it connects the industrial physical space and the network information space, integrating physical entities with communication and computing. This allows the traditional PE chemical production process to achieve a leap forward in intelligent networking. The sensing unit collects data in real time and transmits it to the controller for analysis and processing using wireless network communication technology, enhancing the timeliness and responsiveness of decision-making. The remote controller, located in the cloud, breaks free from the limitations of geographical environment and application scenarios, improves data storage and computing capabilities, and enhances the flexibility and adaptability of decision-making. It can provide more precise control of the production process and ensure the reliable and safe operation of reaction separation. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying any creative work.

[0080] Figure 1 This is a framework diagram of a PE plastic reaction control system based on cyber-physical fusion according to Example 1 of the present invention;

[0081] Figure 2 This is a flow chart of a method for predicting and controlling PE plastic reactions based on cyber-physical fusion under deception attacks according to Example 1 of the present invention;

[0082] Figure 3 This is a structural schematic diagram of the physical process of polymerization reaction and flash separation in Example 1 of the present invention;

[0083] Figure 4 Schematic diagram of the evolution process of the input signal of the deviation system of Example 1 of the present invention;

[0084] Figure 5 Schematic diagram of a model predictive control method for a deviation between an on-site reaction process and a target steady state according to Example 1 of the present invention;

[0085] Figure 6 This is a sequence diagram of random deception attack launch / sleep according to Example 1 of the present invention;

[0086] Figure 7 This is a graph showing the output result of the controlled PE plastic reaction process temperature in Example 1 of the present invention;

[0087] Figure 8 This is a diagram showing the input results of the controlled PE plastic reaction process according to Example 1 of the present invention;

[0088] Figure 9 Box plots showing the mean square error of the controlled temperature under attack compared to normal communication in multiple groups according to Example 1 of the present invention;

[0089] Figure 10 This is a structural block diagram of a PE plastic reaction control device based on cyber-physical fusion according to Example 2 of the present invention;

[0090] Figure 11 This is a structural block diagram of an electronic device according to embodiment 3 of the present invention. DETAILED DESCRIPTION

[0091] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention. It should be understood that the specific embodiments described are only used to explain this application and are not used to limit this application.

[0092] Example 1:

[0093] like Figure 1 As shown, this embodiment provides a PE plastic reaction control system based on cyber-physical fusion, which includes a sensing unit, a control unit, an execution unit, and a communication unit, wherein:

[0094] The sensing unit is used to collect data of the physical entity of the PE plastic industry reaction in real time and convert the collected data into discrete state information; the discrete state information is transmitted to the control unit through the communication unit;

[0095] The control unit is used to store and analyze the received discrete state information to obtain control instructions for regulating the physical entity; and transmit the control instructions to the execution unit through the communication unit;

[0096] The execution unit is used to serialize the received control instructions and adjust the flow rate of the heat-carrying medium in the jacket by the actuator according to the serialized control instructions to achieve heating or cooling of the container, so that the polymerization and separation physical processes of PE plastic production can continue to operate normally;

[0097] The communication unit is used to realize communication between the control unit and the perception unit and execution unit.

[0098] Specifically, the sensing unit in this embodiment includes a sensor and a sampler. The temperature sensor and the concentration sensor are used to collect the container temperature and the concentration of each material in the polymerization reaction and flash separation process, and the information is converted into discrete state information through the sampler.

[0099] The control unit includes a database and a controller, which uses the database to store received status information. The controller is pre-installed with a discrete predictive control algorithm. Based on the received status information, the discrete predictive control algorithm, combined with the target steady-state heat output, calculates the total heat required for the corresponding PE plastic reaction, which serves as the control command signal for regulating the physical entity.

[0100] The execution unit includes a zero-order holder and an actuator. The zero-order holder is used to make the received control instructions continuous. According to the continuous control instructions, the actuator, i.e. the valve of the jacket on the outside of the container, is adjusted to control the flow of the heat-carrying medium in the jacket, thereby heating or cooling the reactor and separator, completing the heat exchange between the container and the jacket, and allowing the polymerization and separation physical processes of PE plastic production to continue to operate normally.

[0101] The communication unit realizes communication between units through the network.

[0102] This embodiment also provides a PE plastic reaction prediction control method based on information-physics fusion, which is used to implement the discrete prediction control algorithm preset in the PE plastic reaction control system controller to solve the control command signal of the regulating physical entity, such as Figure 2 As shown, it mainly includes the following steps:

[0103] S201. Based on the knowledge of PE plastic polymerization and separation mechanism, a deviation system model of the PE plastic reaction physical entity based on the target steady-state point is established based on nonlinear dynamics.

[0104] In the physical process of PE plastic reaction, the PE plastic reaction physical entity includes a flash separator and at least one polymerization reactor.

[0105] Specifically, if Figure 3 As shown, the physical process of the PE plastic reaction in this embodiment includes the structure of two polymerization reactors and a flash separator. The reaction process of PE plastic is as follows: the raw material (marked as A) enters the first reactor, and the PE plastic product (marked as B) is generated through the main reaction, and the product B generates a by-product (marked as C) through a side reaction; then the effluent of the first reactor and another stream of raw material A enter the second reactor, and the same main and side reactions occur in the second reactor; part of the effluent of the second reactor enters the separator and is recycled back to the first reactor. Both the reactor and the separator are equipped with a jacket to heat or cool the container. Based on standard modeling assumptions and material and energy balances, kinetic models are established for the first reactor, the second reactor and the separator respectively.

[0106] The nonlinear dynamic model of the first reactor is:

[0107]

[0108] Among them, C A10 =1, C B10 =0, C A1 is the concentration of raw material A in the first reactor, C B1 is the concentration of product B in the first reactor, T1 is the temperature inside the first reactor, T3 is the temperature inside the separator, C Ar is the concentration of raw material A in the circulating flow, C Br is the concentration of product B in the circulating flow, T 10 is the feed flow temperature of the first reactor, F1 is the discharge flow rate of the first reactor, F 10 is the feed flow rate of the first reactor, F r is the circulation flow rate, V1 is the volume of the first reactor, E1 is the activation energy of the main reaction, E2 is the activation energy of the side reaction, k1 is the pre-exponential value of the main reaction, k2 is the pre-exponential value of the side reaction, ΔH1 is the reaction heat of the main reaction, ΔH2 is the reaction heat of the side reaction, Q1 is the heat exchanged between the first reactor and the jacket, c p is the specific heat capacity, R is the gas constant, and ρ is the solution density.

[0109] The nonlinear dynamic model of the second reactor is:

[0110]

[0111] Among them, C A20 =1, C B20 =0, C A2 is the concentration of raw material A in the second reactor, C B2 is the concentration of product B in the second reactor, T2 is the temperature inside the second reactor, T 20is the feed flow temperature of the second reactor, F 20 is the feed flow rate of the second reactor, V2 is the volume of the second reactor, and Q2 is the heat exchanged between the second reactor and the jacket.

[0112] The nonlinear dynamic model of the separator is:

[0113]

[0114] Among them, C A3 is the concentration of raw material A in the separator, C B3 is the concentration of product B in the separator, T3 is the temperature in the separator, V3 is the volume of the separator, Q3 is the heat exchanged between the separator and the jacket, F p is the clearance flow rate, ΔH vap1 is the vaporization enthalpy of raw material A, ΔH vap2 is the vaporization enthalpy of product B, ΔH vap3 is the vaporization enthalpy of by-product C.

[0115] The concentration of each substance in the circulating flow is related to the separator liquid. The following algebraic equation is established for the circulating flow:

[0116]

[0117] Among them, α A is the relative volatility of raw material A, α B is the relative volatility of product B, α C is the relative volatility of by-product C.

[0118] make is the state variable of the physical entity real system of the PE plastic reaction process, where l=1,2,3,let As the execution input variable of the real system, the first reactor, the second reactor and the separator models (1)-(3) are discretized based on the sampling time of 0.005h, and the discrete nonlinear real system model of the PE plastic reaction process is obtained as follows:

[0119]

[0120] in, is the actual system state variable at the kth sampling moment, The input variables of the real system execution at the kth sampling moment, is a discrete nonlinear function.

[0121] Set the target steady-state point of each container to (C Als ,C Bls ,T ls ,Q ls), l=1,2,3, where C A1s =0.1763 kmol / m 3 , C B1s =0.6731 kmol / m 3 , T 1s =480.3165K,Q 1s =2.9×10 9 kJ / h,C A2s =0.1965 kmol / m 3 , C B2s =0.6536 kmol / m 3 , T 2s =472.7863K,Q 2s =1.0×10 9 kJ / h C A3s =0.0651 kmol / m 3 , C B3s =0.6703 kmol / m 3 , T 3s =474.8877K,Q 3s =2.9×10 9 kJ / h, let is the target steady state, where χ ls =[C Als C Bls T ls ] T , l=1,2,3, let The target steady-state execution input is used. For the real system (5) of the PE plastic reaction process, the linearization processing based on the target steady-state point is performed, and is the state variable of the PE plastic reaction process deviation system, let As the execution input variable of the deviation system, the deviation system model based on the target steady-state point of the PE plastic reaction process is obtained as follows:

[0122]

[0123] Among them, the matrix is the linearized system matrix, the matrix is the linearized input matrix, is a nonlinear function term, There exists a Lipschitz matrix W of appropriate dimension such that the nonlinear term The following Lipschitz conditions are met:

[0124]

[0125] S202. Based on communication security and physical device characteristics, summarize the evolution of the deviation system input signal, combine the deviation system model, and construct an information-physical fusion prediction model for PE plastic reaction.

[0126] Specifically, for a discrete PE plastic reaction process deviation system, a prediction model was constructed within the framework of a discrete cyber-physical system, omitting the sampler and zero-order holder. Within the cyber-physical system framework, the physical and cyber spaces are interconnected, and production and control information is transmitted and evolved in the form of signals.

[0127] like Figure 4 As shown, the two-step evolution of the input signal in the inductive bias system is summarized based on the communication security and the intrinsic characteristics of the actuator.

[0128] The first step of the evolution of the input signal in the deviation system is derived as follows: Let u k The input signal for the controller to calculate the output, is the input signal after transmission through the channel. According to the cyber-physical system framework of discrete processes, the input signal is calculated by the control unit and then transmitted to the execution unit through the network communication link. The actuator adjusts the operation of the PE plastic reaction process deviation system. Therefore, the first step of the input signal evolution occurs before and after the communication, which is

[0129] Considering the risk of external malicious attacks on the communication link connecting the control unit and the execution unit, the attacker launches a spoofing attack in order to destroy the stability of the reaction process and reduce the expected product quality and production efficiency. In the presence of a spoofing attack, the input signal after transmission through the network link is:

[0130]

[0131] Among them, α k is a random variable with the value 0 or 1 describing the link security. k =0 means the link is not attacked, and the signal before and after the communication transmission remains unchanged. When α k =1 indicates that the link has been attacked and the input signal has been tampered with after the communication transmission. where v k Malicious false information injected by the attacker. The deception attack energy is limited, v k ∈Υ, where n u is the dimension of the input vector, ||·|| 2 is the square of the L2 norm, v max is the maximum amplitude vector of a single attack, is the upper bound of the total energy of the attack, so we have is the maximum energy of a single attack,

[0132] For the random attack process, the attack launch is modeled as a random sequence following the Bernoulli distribution. Based on historical attack records, the following prior probability model of deception attack is obtained:

[0133]

[0134] Among them, Prob{·} represents probability, is the attack probability determined based on past attack records.

[0135] The second step evolution of the input signal in the deviation system is derived from the following process: Based on the discrete deviation system model of the PE plastic reaction process, is the actual input signal for the deviation system. Due to the inherent opening limit of the jacket regulating valve, the heat transfer medium in the jacket can only provide or remove a limited amount of heat to the container, so the actuator has nonlinear saturation characteristics. Therefore, the second step of the input signal evolution occurs before and after the execution, which is

[0136] For the unit saturation limit, the actuator receives an input signal transmitted via the link After that, the actual execution input for:

[0137]

[0138] Where sat(·) is the standard unit saturation function under scalar or vector, that is,

[0139] is a vector The rth element of .

[0140] Considering the attacker of the deception attack has the knowledge of the saturation level of the actuator, in order to save attack energy, the attack signal applied each time is within the saturation range. Therefore, combining the first step evolution (8) and the second step evolution (10) of the input signal, the actual execution input of the deviation system is restated as follows: k The expression:

[0141]

[0142] Combining the physical entity discrete deviation system model (6) and the input signal evolution relationship (11) in step S201, the following information-physics fusion prediction model for PE plastic reaction deviation is obtained:

[0143]

[0144] S203. Based on the information-physical fusion prediction model and prior deception attack knowledge, design performance indicators and control laws, and establish minimization control optimization problems and constraints.

[0145] like Figure 5 As shown, based on the information-physics fusion prediction model of PE plastic reaction deviation, the controlled output for control performance is constructed. is the state weight matrix, is the input weight matrix.

[0146] Since the container ambient temperature seriously affects the polymerization and separation of PE plastic, the temperature must be quickly reached and stably maintained at the target steady-state point. Therefore, the weight matrix of the controlled output is selected as I is the identity matrix of appropriate dimension.

[0147] Design a predictive control algorithm to solve the following minimum and maximum performance objectives for each sampling time k:

[0148]

[0149] Among them, the infinite time domain mean square performance index is represents the mathematical expectation, z k+i|k The system is controlled to output the predicted value at the k+ith moment in the future under the kth sampling moment.

[0150] For the control method, the control law adopts the linear state feedback control law u k+i|k =F k x k+i|k , where F k is the state feedback control gain at the kth sampling moment, x k+i|k The predicted value of the system state at the kth sampling moment in the future k+i moment, u k+i|k The predicted value of the input signal output by the system controller at the kth sampling moment in the future k+i-th moment.

[0151] Furthermore, based on the convex representation method, the nonlinear problem of input saturation in the cyber-physical fusion prediction model is solved, including:

[0152] For the non-unit saturation limit of the PE plastic reaction information physical fusion prediction model, the following unitization processing is performed:

[0153]

[0154] in, U=diag{u max}, diag{·} represents the diagonal block matrix, u max is the maximum input saturation level, in this embodiment, umax =[2.9×10 9 1.0×10 9 2.9×10 9 ] T For simplicity, the subsequent superscript ∧ is omitted.

[0155] Introducing the auxiliary state feedback controller H k x k+i|k , where H k is the auxiliary controller gain, when for all x k+i|k , can satisfy When the saturation auxiliary condition is met, the saturated input signal sat(u k+i|k )=sat(F k x k+i|k ) can be restated as the following convex representation:

[0156]

[0157] Where 0≤η j ≤1, E j is a diagonal matrix with diagonal elements of 0 or 1, Represents a symmetric polyhedron Represents the matrix H k The rth row of .

[0158] Furthermore, based on the Lyapunov function method, the mean square safety problem of the predictive control algorithm is solved, including:

[0159] Set up a quadratic Lyapunov function Matrix P k > 0. At sampling time k, assume that the predicted value x for all future k+i moments is k+i|k 、u k+i|k and v k+i|k , function V(x k+i|k ) can satisfy the following input-state stability constraints in the mean square sense:

[0160]

[0161] Among them, λ>0 is the mean square security level, which indicates the ability to resist attack signals. The smaller λ is, the better the ability is. ∞|k =0, so V(x ∞|k )=0.

[0162] Add both ends of equation (14) from i = 0 to i = ∞, and we get the objective function J k An upper bound of Therefore, the model predictive control method under deception attack is restated as a minimization control optimization problem, that is, to design a state feedback control law u at each sampling time k k+i|k =F k x k+i|k To minimize the performance upper bound, a non-negative scalar γ is introduced k So that:

[0163] V(x k|k )≤γ k (15)

[0164] The upper bound of the target performance is Since the parameter λ is adjustable, φ is introduced k =λ 2 γ k , so that the objective of the minimization control optimization problem can be restated as minimizing the parameter φ k .

[0165] Furthermore, based on the quadratic boundedness technique, the invariant set problem of the predictive control algorithm is solved, including:

[0166] make Setting Collection At sampling time k, assume that for all x k+i|k 、u k+i|k and v k+i|k , can be in When deduced Then for any future time Therefore, the saturation auxiliary condition was restated as

[0167] Furthermore, the predictive control method can be restated as a minimization control optimization problem: Based on the PE plastic reaction deviation state information x received at sampling time k k and the cyber-physical fusion prediction model (12), given the adjustable parameters ε1>0, ε2>0, 0<β<1, if there exists a matrix Q k >0, G k 、Y k , Z k and the scalar γ k >0,δ k >0,φ k > 0 so that the following minimization problem is Solvable:

[0168]

[0169] Security constraints:

[0170]

[0171] Q k ≥δ k I(19)

[0172] Quadratic boundedness conditions:

[0173]

[0174] State constraints:

[0175]

[0176] Input saturation auxiliary constraints:

[0177]

[0178] in, Y k =F k G k , Z k =H k G k , X k,nn is the matrix X k The nth diagonal element of max,n is the absolute upper limit of the nth state, is the matrix Z k The rth row of , 0 is the zero matrix of the corresponding dimension, n x is the dimension of the state variable.

[0179] S204. Based on the real-time reaction state information, rolling calculation is performed to minimize the control optimization problem, and the control instructions of the physical entity are updated and outputted in combination with the target production steady-state heat to achieve safe predictive control of the PE plastic reaction process.

[0180] At each sampling moment, based on the received real-time reaction control system state information, the control optimization problem (16)-(22) is minimized and calculated according to Design control law F k x k ; Then pass The non-unit saturation reduction process is: (Fkxk)r represents the rth element of the control law Fkxk result, sat((F k x k ) r )=sign((F k x k ) r )min{1,|(F k x k ) r |}, r=1,2,…,n u; Finally, the required container heat input is combined with the target production steady state but The control instructions for regulating the physical entity output by the control unit in the PE plastic reaction control system, namely the heating / cooling control instructions of each container, are output to control the jacket regulating valve actuator of each container to achieve temperature control of the polymerization reaction and flash separation, so that the physical entity of the PE plastic reaction process tends to the production target steady state.

[0181] Random attack sequence and final control result are as follows Figures 6-9 shown. Figure 6 This is a possible random deception attack launch / sleep sequence diagram; Figure 7 This is the output result diagram of the temperature of the controlled PE plastic reaction process; Figure 8 It is the input result diagram of the controlled PE plastic reaction process; Figure 9 The following are box plots of the mean square error of the controlled temperature under attack compared to normal communication. The simulation results show that when the network link connecting the control unit and the execution unit in the cyber-physical fusion PE plastic reaction control system is subjected to a random deception attack, the secure model predictive control algorithm (secure MPC) can more quickly guide the container ambient temperature to the expected equilibrium steady-state point, achieving a temperature control effect closer to that under normal communication conditions. This improves the cyber-physical system's resistance to external malicious attacks and its safety performance, and has a positive effect on improving the quality of PE plastic products and ensuring the reliable operation of the production process.

[0182] The method provided in this embodiment can free the traditional PE plastic production industry from the limitations of the actual environment and hardware, provide more accurate, flexible and fast real-time perception and dynamic control for the production process, and ensure the reliable and safe operation of PE plastic reaction and separation.

[0183] Those skilled in the art will appreciate that all or part of the steps in the method for implementing the above embodiments may be completed by instructing related hardware through a program, and the corresponding program may be stored in a computer-readable storage medium.

[0184] It should be noted that although the method operations of the above embodiments are described in a particular order in the accompanying drawings, this does not require or imply that the operations must be performed in this particular order, or that all of the illustrated operations must be performed to achieve the desired results. Rather, the depicted steps may be performed in a different order. Additionally or alternatively, certain steps may be omitted, multiple steps may be combined into a single step, and / or a single step may be broken down into multiple steps.

[0185] Example 2:

[0186] like Figure 10As shown, this embodiment provides a PE plastic reaction prediction and control device based on cyber-physical fusion, which includes a first building block 1001, a second building block 1002, a third building block 1003 and a calculation module 1004, wherein:

[0187] The first building block 1001 is used to establish a deviation system model of the PE plastic reaction physical entity based on a target steady-state point based on nonlinear dynamics according to the knowledge of the polymerization and separation mechanism of PE plastic;

[0188] The second construction module 1002 is used to summarize the evolution of the input signal in the deviation system based on communication security and physical device characteristics, and to construct an information-physical fusion prediction model for PE plastic reaction deviation in combination with the deviation system model;

[0189] The third building block 1003 is used to design performance indicators and control laws based on the cyber-physical fusion prediction model and prior deception attack knowledge, and to establish a minimization control optimization problem and constraints;

[0190] The calculation module 1004 is used to calculate the minimization control optimization problem based on the real-time discrete state information, and obtain the control instructions for regulating the physical entity in combination with the target production steady-state heat to achieve safe predictive control of the PE plastic reaction process.

[0191] The specific implementation of each module in this embodiment can be found in the above-mentioned embodiment 1, and will not be described one by one here; it should be noted that the device provided in this embodiment is only illustrated by the division of the above-mentioned functional modules. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure can be divided into different functional modules to complete all or part of the functions described above.

[0192] Example 3:

[0193] This embodiment provides an electronic device, which may be a terminal. Figure 11As shown, it comprises a processor 1102, a memory, an input device 1103, a display 1104 and a network interface 1105 connected via a system bus 1101. The processor is used to provide computing and control capabilities. The memory includes a non-volatile storage medium 1106 and an internal memory 1107. The non-volatile storage medium 1106 stores an operating system, a computer program and a database. The internal memory 1107 provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. When the processor 1102 executes the computer program stored in the memory, the PE plastic reaction prediction control method based on cyber-physical fusion of the above-mentioned embodiment 1 is implemented, that is, according to the PE Based on the knowledge of plastic polymerization and separation mechanism, a deviation system model based on the target steady-state point of the PE plastic reaction physical entity is established based on nonlinear dynamics; based on communication security and physical device characteristics, the evolution of the input signal in the deviation system is summarized, and combined with the deviation system model, an information-physical fusion prediction model of the PE plastic reaction deviation is constructed; based on the information-physical fusion prediction model and prior deception attack knowledge, performance indicators and control laws are designed, and the minimization control optimization problem and constraints are established; the minimization control optimization problem is calculated according to the real-time discrete state information, and the control instructions for regulating the physical entity are obtained in combination with the target production steady-state heat to achieve safe predictive control of the PE plastic reaction process.

[0194] Example 4:

[0195] This embodiment provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the PE plastic reaction prediction and control method based on information-physical fusion of the above-mentioned embodiment 1, namely: according to the knowledge of PE plastic polymerization and separation mechanism, a deviation system model based on the target steady-state point of the PE plastic reaction physical entity is established based on nonlinear dynamics; based on communication security and physical device characteristics, the evolution of the input signal in the deviation system is summarized, and combined with the deviation system model, an information-physical fusion prediction model of the PE plastic reaction deviation is constructed; based on the information-physical fusion prediction model and prior deception attack knowledge, performance indicators and control laws are designed, and a minimization control optimization problem and constraints are established; the minimization control optimization problem is calculated based on real-time discrete state information, and the control instructions for regulating the physical entity are obtained in combination with the target production steady-state heat to achieve safe predictive control of the PE plastic reaction process.

[0196] It should be noted that the computer-readable storage medium of the present embodiment may be a computer-readable signal medium or a computer-readable storage medium or any combination thereof. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof.

[0197] The above is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes based on the technical solution and inventive concept of the present invention within the scope disclosed by the present invention, which falls within the scope of protection of the present invention.

Claims

1. A PE plastic reaction prediction and control method based on information-physics fusion, characterized in that: The method comprises: According to the knowledge of PE plastic polymerization and separation mechanism, a deviation system model based on the target steady-state point of the PE plastic reaction physical entity is established based on nonlinear dynamics; Based on communication security and physical device characteristics, the evolution of input signals in the deviation system is summarized, and combined with the deviation system model, an information-physical fusion prediction model for PE plastic reaction deviation is constructed; Based on the cyber-physical fusion prediction model and prior knowledge of deception attacks, performance indicators and control laws are designed, and the minimization control optimization problem and constraints are established; Based on the real-time discrete state information, the minimization control optimization problem is calculated and combined with the target production steady-state heat to obtain the control instructions for regulating the physical entity to achieve safe predictive control of the PE plastic reaction process; The design of performance indicators and control laws based on the cyber-physical fusion prediction model and prior deception attack knowledge, and the establishment of minimization control optimization problems and constraints, include: Based on the cyber-physical fusion prediction model, a controlled output for controlling performance is constructed. is the state weight matrix, is the input weight matrix; x k is the state variable of the PE plastic reaction process deviation system at the kth sampling moment, u k is the input signal of the deviation system at the kth sampling moment; sat(·) is the standard unit saturation function under scalar or vector; The performance objective of solving the minimum maximization of the following expression for each sampling time k is: Where, the infinite time domain mean square performance index is represents the mathematical expectation, z k+i|k is the predicted value of the system controlled output at the kth sampling moment in the future k+i moment, and the control law adopts the linear state feedback control law u k+i|k =F k x k+i|k , F k is the state feedback control gain at the kth sampling moment, x k+i|k is the predicted value of the system state at the kth sampling moment in the future k+i moment, u k+i|k is the predicted value of the input signal output by the system controller at the k+ith moment in the future under the kth sampling moment; The minimization control optimization problem and constraints are: Given adjustable parameters ε1>0, ε2>0, 0<β<1, there exists a matrix Q k >0, G k 、Y k , Z k and the scalar γ k >0,δ k >0,φ k > 0 so that the following minimization problem is Solvable: Security constraints: Quadratic boundedness conditions: State constraints: Q k ≤X k , n=1,2,…,n x Input saturation auxiliary constraints: Where, is the system matrix, is the input matrix; is the attack probability; Y k =F k G k , Z k =H k G k , X k,nn is the matrix X k The nth diagonal element of max,n is the absolute upper limit of the nth state, is the matrix Z k The rth row of , 0 is the zero matrix of the corresponding dimension, I is the identity matrix of the corresponding dimension, W is the Lipschitz matrix, E j is a diagonal matrix with diagonal elements of 0 or 1, n x is the dimension of the state variable, n u is the dimension of the input variables, is the maximum energy of a single attack; Obtain the feedback control gain.

2. The PE plastic reaction prediction and control method according to claim 1, characterized in that: The information-physics fusion prediction model for PE plastic reaction deviation is: Where x k+1 、x k are the state variables of the cyber-physical fusion system of PE plastic reaction deviation at the k+1th and kth sampling moments respectively; is the system matrix; is the input matrix; α k is a random variable describing the link security; v k is the malicious false information injected by the attacker; sat(·) is the standard unit saturation function under scalar or vector; u k is the input signal at the kth sampling moment; g(·,·) is a nonlinear function.

3. The PE plastic reaction prediction and control method according to claim 1, characterized in that: Based on the communication security and physical device characteristics, the evolution of the input signal in the deviation system is summarized, including the evolution of the input signal before and after communication and the evolution before and after the actuator is executed in two steps. The input signal actually executed by the actuator of the deviation system is obtained by summarizing the two-step evolution; The evolution of the input signal before and after communication is specifically: For the deviation system, let u k is the input signal output by the controller, for u k After the input signal is transmitted by the communication unit, then: Where, α k =0 means the link is not attacked, α k =1 indicates that the link has been attacked; v k ∈Υ, n u is the dimension of the input vector, ||·|| 2 is the square of the L2 norm, v max is the maximum amplitude vector of a single attack, is the upper bound of the total energy of the attack, so The evolution of the executor before and after execution is as follows: make Received by the actuator The input signal actually executed after is the execution input variable of the deviation system at the kth sampling moment, then: where sat(·) is the standard unit saturation function under scalar or vector.

4. The PE plastic reaction prediction and control method according to claim 3, characterized in that: For the random attack process, the attack launch is modeled as a random sequence following the Bernoulli distribution. According to the historical attack records, the prior probability model is obtained: Where Prob{·} represents probability; The value of is determined based on past attack records.

5. The PE plastic reaction prediction and control method according to claim 1, characterized in that: According to the knowledge of PE plastic polymerization and separation mechanism, a deviation system model based on the target steady-state point of the PE plastic reaction physical entity is established based on nonlinear dynamics, including: make is the state variable of the PE plastic reaction physical entity real system at the kth sampling moment, is the execution input variable of the PE plastic reaction physical entity real system at the kth sampling moment; is the target steady-state state of the physical entity, Input for the target steady-state execution of the physical body; As the state variable of the PE plastic reaction process deviation system at the kth sampling moment, is the execution input variable of the deviation system at the kth sampling moment; The discrete nonlinear real system model of the PE plastic reaction process is linearized based on the target steady-state point, and the deviation system model of the PE plastic reaction process based on the target steady-state point is obtained as follows: Where x k+1 is the state variable of the PE plastic reaction process deviation system at the k+1th sampling time; is the linearized system matrix, is the linearized input matrix, It is the discrete nonlinear system function of the PE plastic reaction physical entity real system; It is the nonlinear function term of the deviation system of PE plastic reaction process.

6. The PE plastic reaction prediction and control method according to claim 5, characterized in that: The PE plastic reaction physical entity includes a flash separator and at least one polymerization reactor; The discrete nonlinear real system model is obtained through the following process: make is the state variable of the physical entity real system of the PE plastic reaction process, Input variables for the execution of physical reality systems; The nonlinear dynamic model of the PE plastic reaction physical entity is discretized based on sampling time, and the discrete nonlinear real system model of the PE plastic reaction process is obtained as follows: Where, is the state variable at the k+1th sampling moment, is a discrete nonlinear function.

7. A PE plastic reaction control system based on information-physics fusion, characterized in that: The system includes a sensing unit, a control unit, an execution unit, and a communication unit, wherein: The sensing unit is used to collect data of the physical entity of the PE plastic industry reaction in real time and convert the collected data into discrete state information; the discrete state information is transmitted to the control unit through the communication unit; A control unit, configured to store the received discrete state information and analyze the PE plastic reaction prediction control method according to any one of claims 1 to 6 to obtain a control instruction for adjusting the physical entity; and transmit the control instruction to the execution unit via the communication unit; the control instruction for adjusting the physical entity being the total heat required for the PE plastic reaction; The execution unit is used to serialize the received control instructions and adjust the flow rate of the heat-carrying medium in the jacket by the actuator according to the serialized control instructions to achieve heating or cooling of the container, so that the polymerization and separation physical processes of PE plastic production can continue to operate normally; The communication unit is used to realize communication between the control unit and the perception unit and execution unit.

8. The PE plastic reaction control system according to claim 7, characterized in that: The collected data include the container temperature and the concentration of each material in the polymerization reaction and flash separation process.

9. A PE plastic reaction prediction and control device based on information-physics fusion, characterized in that: The device comprises: The first building block is used to establish a deviation system model of the PE plastic reaction physical entity based on the target steady-state point based on nonlinear dynamics according to the knowledge of PE plastic polymerization and separation mechanism; The second building block is used to summarize the evolution of input signals in the deviation system based on communication security and physical device characteristics, and to construct an information-physical fusion prediction model for PE plastic reaction deviation in combination with the deviation system model; The third building block is used to design performance indicators and control laws based on the cyber-physical fusion prediction model and prior deception attack knowledge, and to establish the minimization control optimization problem and constraints; The calculation module is used to calculate the minimization control optimization problem based on real-time discrete state information and obtain the control instructions for regulating the physical entity in combination with the target production steady-state heat to achieve safe predictive control of the PE plastic reaction process; The third building block is specifically used for: Based on the cyber-physical fusion prediction model, a controlled output for controlling performance is constructed. is the state weight matrix, is the input weight matrix; x k is the state variable of the PE plastic reaction process deviation system at the kth sampling moment, u k is the input signal of the deviation system at the kth sampling moment; sat(·) is the standard unit saturation function under scalar or vector; The performance objective of solving the minimum maximization of the following expression for each sampling time k is: Where, the infinite time domain mean square performance index is represents the mathematical expectation, z k+i|k is the predicted value of the system controlled output at the kth sampling moment in the future k+i moment, and the control law adopts the linear state feedback control law u k+i|k =F k x k+i|k , F k is the state feedback control gain at the kth sampling moment, x k+i|k is the predicted value of the system state at the kth sampling moment in the future k+i moment, u k+i|k is the predicted value of the input signal output by the system controller at the k+ith moment in the future under the kth sampling moment; The minimization control optimization problem and constraints are: Given adjustable parameters ε1>0, ε2>0, 0<β<1, there exists a matrix Q k >0, G k 、Y k , Z k and the scalar γ k >0,δ k >0,φ k > 0 so that the following minimization problem is Solvable: Security constraints: Quadratic boundedness conditions: State constraints: Q k ≤X k , n=1,2,…,n x Input saturation auxiliary constraints: Where, is the system matrix, is the input matrix; is the attack probability; Y k =F k G k , Z k =H k G k , X k,nn is the matrix X k The nth diagonal element of max,n is the absolute upper limit of the nth state, is the matrix Z k The rth row of , 0 is the zero matrix of the corresponding dimension, I is the identity matrix of the corresponding dimension, W is the Lipschitz matrix, E j is a diagonal matrix with diagonal elements of 0 or 1, n x is the dimension of the state variable, n u is the dimension of the input variables, is the maximum energy of a single attack; Obtain the feedback control gain.

Images