CN117588394A - AIoT-based intelligent linkage control method and system for vacuum pump - Google Patents
AIoT-based intelligent linkage control method and system for vacuum pump Download PDFInfo
- Publication number
- CN117588394A CN117588394A CN202410075811.1A CN202410075811A CN117588394A CN 117588394 A CN117588394 A CN 117588394A CN 202410075811 A CN202410075811 A CN 202410075811A CN 117588394 A CN117588394 A CN 117588394A
- Authority
- CN
- China
- Prior art keywords
- vacuum pump
- control strategy
- representing
- control
- state
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 142
- 238000011217 control strategy Methods 0.000 claims abstract description 96
- 230000008569 process Effects 0.000 claims abstract description 78
- 238000013528 artificial neural network Methods 0.000 claims abstract description 44
- 238000012544 monitoring process Methods 0.000 claims abstract description 19
- 238000012549 training Methods 0.000 claims abstract description 5
- 230000006870 function Effects 0.000 claims description 45
- 230000009471 action Effects 0.000 claims description 29
- 230000007613 environmental effect Effects 0.000 claims description 24
- 238000007781 pre-processing Methods 0.000 claims description 15
- 239000011148 porous material Substances 0.000 claims description 14
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 14
- 239000003673 groundwater Substances 0.000 claims description 13
- 238000004062 sedimentation Methods 0.000 claims description 13
- 230000007704 transition Effects 0.000 claims description 9
- 102100040653 Tryptophan 2,3-dioxygenase Human genes 0.000 claims description 6
- 101710136122 Tryptophan 2,3-dioxygenase Proteins 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 6
- 230000005540 biological transmission Effects 0.000 claims description 6
- 238000004891 communication Methods 0.000 claims description 6
- 230000000875 corresponding effect Effects 0.000 claims description 6
- 238000005265 energy consumption Methods 0.000 claims description 6
- 238000005457 optimization Methods 0.000 claims description 6
- 238000013461 design Methods 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 238000011282 treatment Methods 0.000 abstract description 20
- 230000002787 reinforcement Effects 0.000 abstract description 6
- 230000008859 change Effects 0.000 abstract description 4
- 238000013507 mapping Methods 0.000 abstract description 3
- 230000000694 effects Effects 0.000 description 7
- 238000007789 sealing Methods 0.000 description 5
- 238000010276 construction Methods 0.000 description 4
- 230000010354 integration Effects 0.000 description 4
- 230000005611 electricity Effects 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 238000011369 optimal treatment Methods 0.000 description 3
- 239000004576 sand Substances 0.000 description 3
- 239000002689 soil Substances 0.000 description 3
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003825 pressing Methods 0.000 description 1
- 238000005086 pumping Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- E—FIXED CONSTRUCTIONS
- E02—HYDRAULIC ENGINEERING; FOUNDATIONS; SOIL SHIFTING
- E02D—FOUNDATIONS; EXCAVATIONS; EMBANKMENTS; UNDERGROUND OR UNDERWATER STRUCTURES
- E02D3/00—Improving or preserving soil or rock, e.g. preserving permafrost soil
- E02D3/02—Improving by compacting
- E02D3/10—Improving by compacting by watering, draining, de-aerating or blasting, e.g. by installing sand or wick drains
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F04—POSITIVE - DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS FOR LIQUIDS OR ELASTIC FLUIDS
- F04B—POSITIVE-DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS
- F04B49/00—Control, e.g. of pump delivery, or pump pressure of, or safety measures for, machines, pumps, or pumping installations, not otherwise provided for, or of interest apart from, groups F04B1/00 - F04B47/00
- F04B49/02—Stopping, starting, unloading or idling control
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F04—POSITIVE - DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS FOR LIQUIDS OR ELASTIC FLUIDS
- F04B—POSITIVE-DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS
- F04B49/00—Control, e.g. of pump delivery, or pump pressure of, or safety measures for, machines, pumps, or pumping installations, not otherwise provided for, or of interest apart from, groups F04B1/00 - F04B47/00
- F04B49/06—Control using electricity
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F04—POSITIVE - DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS FOR LIQUIDS OR ELASTIC FLUIDS
- F04B—POSITIVE-DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS
- F04B49/00—Control, e.g. of pump delivery, or pump pressure of, or safety measures for, machines, pumps, or pumping installations, not otherwise provided for, or of interest apart from, groups F04B1/00 - F04B47/00
- F04B49/20—Control, e.g. of pump delivery, or pump pressure of, or safety measures for, machines, pumps, or pumping installations, not otherwise provided for, or of interest apart from, groups F04B1/00 - F04B47/00 by changing the driving speed
-
- F—MECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
- F04—POSITIVE - DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS FOR LIQUIDS OR ELASTIC FLUIDS
- F04B—POSITIVE-DISPLACEMENT MACHINES FOR LIQUIDS; PUMPS
- F04B49/00—Control, e.g. of pump delivery, or pump pressure of, or safety measures for, machines, pumps, or pumping installations, not otherwise provided for, or of interest apart from, groups F04B1/00 - F04B47/00
- F04B49/22—Control, e.g. of pump delivery, or pump pressure of, or safety measures for, machines, pumps, or pumping installations, not otherwise provided for, or of interest apart from, groups F04B1/00 - F04B47/00 by means of valves
-
- 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
- Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Mechanical Engineering (AREA)
- Structural Engineering (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Soil Sciences (AREA)
- Mining & Mineral Resources (AREA)
- Paleontology (AREA)
- Civil Engineering (AREA)
- Environmental & Geological Engineering (AREA)
- Agronomy & Crop Science (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses an AIoT-based intelligent linkage control method and system for a vacuum pump. According to the scheme, a sensor is used for monitoring foundation environment conditions, the foundation environment conditions are described through a Markov decision process, and a nonlinear mapping relation between complex foundation environment conditions and a vacuum pump control strategy is learned; the control strategy is generated by constructing the deep neural network, training the neural network by adopting the reinforcement learning algorithm, introducing trust zone constraint to limit the strategy change of each update, avoiding excessive update of the control strategy, increasing the flexibility of the control of the vacuum pump in the foundation treatment and improving the efficiency of the foundation treatment process.
Description
Technical Field
The invention relates to the technical field of foundation treatment, in particular to an AIoT-based intelligent linkage control method and system for a vacuum pump.
Background
At present, a vacuum preloading method is often adopted in the reinforcement treatment of a large-area high-water-content soft soil foundation, wherein the vacuum preloading method is a method that a plastic drain board is arranged in a soil body to serve as a vertical drain body, a sand cushion layer with a certain thickness is paved on the surface of the foundation, a drain filter tube is paved in the sand cushion layer, an airtight sealing film is paved on the sand cushion layer, a sealing wall is arranged on the boundary, then a vacuum pump is used for pumping air, vacuum negative pressure is generated under the film, a pressure difference is formed between the inside and the outside of the film under the action of atmospheric pressure, pore water in the soil body is discharged through a drain system under the action of the pressure difference, and finally the reinforcement method of the consolidated and settled foundation is achieved.
In the vacuum pre-pressing foundation treatment, a vacuum pump plays a role in vacuumizing, process conditions can be adjusted by using the vacuum pump to adapt to different treatment requirements, smooth process is ensured, parameters are manually set by traditional vacuum pump control, and the parameters of the vacuum pump are required to be adjusted based on experience and expert knowledge so as to achieve the required vacuum level and the optimal treatment effect, so that the maintenance of the optimal performance becomes very complex when facing different working conditions, and the adjustment is difficult in time in a complex environment; the control strategy of the general vacuum pump is influenced by the foundation treatment effect during operation, including foundation vacuum pressure, sedimentation, groundwater level and pore water pressure, and in the whole foundation treatment process, the vacuum pump is always in a working state, even when the vacuum degree under a sealing film reaches a certain level, the vacuum pump still continuously works when the design requirement is met, and the problem of wasting construction electricity is solved.
Disclosure of Invention
Aiming at the situation, in order to overcome the defects of the prior art, the invention provides the intelligent linkage control method and the system for the vacuum pump based on AIoT, aiming at the problems that the vacuum pump parameter needs to be adjusted based on experience and expertise to achieve the required vacuum level and the optimal treatment effect by manually setting the parameter in the traditional vacuum pump control, the maintenance of the optimal performance becomes very complex when facing different working conditions and is difficult to adjust in time in the complex environment, the scheme uses a sensor to monitor the foundation environmental condition, describes the foundation environmental condition through a Markov decision process, learns the nonlinear mapping relation between the more complex foundation environmental condition and the vacuum pump control strategy, and thus the working state of the vacuum pump is adjusted more intelligently; aiming at the influence of the foundation treatment effect when a common vacuum pump control strategy operates, including foundation vacuum pressure, sedimentation, groundwater level and pore water pressure, and in the whole foundation treatment process, the vacuum pump is always in a working state, even when the vacuum degree under a sealing film reaches a certain level, the vacuum pump still continuously works, the problem of wasting construction electricity is solved, the deep neural network is constructed to generate the control strategy, the neural network is trained by adopting a reinforcement learning algorithm, the strategy change updated each time is limited by introducing trust area constraint, the stability of the algorithm is improved, the excessive update of the control strategy is avoided, the flexibility of the vacuum pump control in the foundation treatment is increased, and the efficiency of the foundation treatment process is improved.
The technical scheme adopted by the invention is as follows: the invention provides an AIoT-based intelligent linkage control method for a vacuum pump, which comprises the following steps:
step S1: the method comprises the steps of data acquisition and preprocessing, deploying a sensor to monitor foundation environmental conditions including vacuum pressure, sedimentation, groundwater level and pore water pressure, transmitting the foundation environmental conditions acquired by the sensor to a control center by using an Internet of things communication protocol, and receiving the foundation environmental conditions transmitted by the sensor by the control center and analyzing and preprocessing;
step S2: initializing a control strategy of the vacuum pump by using a Markov decision process, wherein the Markov decision process comprises a state space, an action space, a state transition probability, a reward function and a discount factor of the vacuum pump;
step S3: the control strategy is updated, a deep neural network for representing the control strategy is constructed, a state space of a vacuum pump is received as input, the probability of executing each action in a given state is output, the deep neural network is trained by using a TRPO algorithm, parameters of the deep neural network are iteratively optimized, and a final control strategy is generated;
step S4: and (3) remote monitoring and control, namely deploying a final control strategy to a control center, generating a control instruction, transmitting the control instruction to a vacuum pump and other related equipment through an Internet of things transmission protocol, and realizing real-time adjustment of the vacuum pump through remote monitoring.
Further, in step S2, the control strategy is initialized, which specifically includes the following steps:
step S21: defining a control strategy, wherein the aim of defining the control strategy is to maximize the operation efficiency of the vacuum pump, minimize the energy consumption of the vacuum pump, and the constraint of the control strategy is the operation range of the vacuum pump;
step S22: establishing an MDP model, and modeling a control strategy of the vacuum pump through a Markov decision process, wherein the Markov decision process comprises a state space, an action space, a state transition probability, a reward function and a discount factor of the vacuum pump, and the following formula is adopted:
；
in the method, in the process of the invention,representing a Markov decision process,/->Representing state space, ++>Representing the action space->Representing state transition probabilities>Representing a reward function->Representing a discount factor;
step S23: defining MDP parameters, wherein the state space of the MDP comprises the running state of a vacuum pump, the vacuum level, sedimentation, the ground water level and pore water pressure, the action space of the MDP comprises the rotating speed of the vacuum pump, the opening degree of a valve and start-stop control, defining a reward function, setting a discount factor to be 0.7 by using a linear weighted summation method, and adopting the following formula:
；
in the method, in the process of the invention,rewarding the operating efficiency of the vacuum pump, +.>Rewarding the energy consumption of the vacuum pump, +.>Rewarding the operating range of the vacuum pump, +.>、/>、/>Weighting corresponding rewarding items;
step S24: defining a jackpot, maximizing the desired prize by selecting an optimal strategy, using the formula:
；
in the method, in the process of the invention,representing an optimal strategy->Representing policy ++>Is (are) desirable to be (are)>Representation->Status of moment->Representation->Action at time.
Further, in step S3, the control strategy updating specifically includes the following steps:
step S31: the neural network design, construct the deep neural network used for representing the control strategy, the deep neural network is made up of input layer, hidden layer, output layer, the state space of MDP of input layer input, hidden layer uses tanh to activate the function to learn and represent complex functional relation from the input data, the formula used is as follows:
；
in the method, in the process of the invention,state information representing an input;
the output layer outputs the probability of performing the corresponding action in the given state, using the softmax function as the activation function of the output layer, using the following formula:
；
in the method, in the process of the invention,representing the +.>Element(s)>Representing the length of the input vector;
step S32: forward propagation, the forward propagation process of the deep neural network is constructed, and the following formula is used:
；
in the method, in the process of the invention,representing an activation value +.>Representing a weight matrix, +.>Representing the bias;
step S33: the objective function is designed to maximize the desired jackpot, using the following formula:
；
in the method, in the process of the invention,is a parameter of the policy, ++>Is a trace from an initial state to a termination state, +.>Is policy, & lt>Is at the moment +.>Rewards obtained at the department, ->Is a discount factor;
step S34: calculating a dominance estimate, estimating the relative superiority of each state-action pair using the following formula:
；
in the method, in the process of the invention,is a state-action pair->Is advantaged by->State-action pair->Is a function of the value of (c),is state->Is a function of the value of (2);
step S35: defining a loss function, and updating the direction of the parameters of the deep neural network by using a strategy gradient loss, wherein the following formula is used:
；
in the method, in the process of the invention,representing policy(s)>Representing +.>According to policy->Is subjected to a calculation of the expected value of the distribution of (c),representation strategy->Middle action->In state->The logarithmic probability of the lower case is related to the parameter->Is a gradient of (2);
step S36: selecting an optimization algorithm, and updating parameters of a strategy network by using a gradient ascent method to increase an objective function value, wherein the following formula is adopted:
；
in the method, in the process of the invention,is->Model parameters for the next iteration->Is learning rate (I/O)>Is the gradient of the loss function to the model parameters;
step S37: the trust zone is calculated, the KL divergence constraint control strategy is used for updating the trust zone of the amplitude, and the following formula is used:
；
in the method, in the process of the invention,indicating that the old policy is in the given state +.>The probability distribution of actions->Representing the probability distribution of actions of the current strategy in the same state,/-, for example>Representing +.>The integration is performed and the integration is performed,is a preset threshold value;
step S38: optimizing the objective function, and maximizing the objective function in the trust zone, wherein the formula is as follows:
；
in the method, in the process of the invention,is an optimized objective function;
step S39: and generating a control strategy, and adjusting parameters of the deep neural network through training optimization to generate a final control strategy.
The invention provides an AIoT-based intelligent linkage control system for a vacuum pump, which comprises a data acquisition and preprocessing module, a control strategy initialization module, a control strategy updating module and a remote monitoring and control module;
the data acquisition and preprocessing module deploys a sensor to monitor foundation environmental conditions including vacuum pressure, sedimentation, groundwater level and pore water pressure, the foundation environmental conditions acquired by the sensor are transmitted to a control center by using an Internet of things communication protocol, the control center receives the foundation environmental conditions transmitted by the sensor and analyzes and preprocesses the foundation environmental conditions to obtain preprocessed data, and the preprocessed data is transmitted to the control strategy initialization module;
the control strategy initializing module receives data from the data acquisition and preprocessing module, defines a Markov decision process by using the received data, and sends the Markov decision process to the control strategy updating module;
the control strategy updating module receives a Markov decision process from the control strategy initializing module, constructs a deep neural network for representing the control strategy, receives the state of the vacuum pump as input, outputs the probability of executing each action in a given state, trains the deep neural network by using a TRPO algorithm, iteratively optimizes parameters of the deep neural network, generates a final control strategy, and sends the final control strategy to the remote monitoring and control module;
the remote monitoring and control module receives data from the control strategy updating module, deploys the final control strategy to the control center, generates control instructions, transmits the control instructions to the vacuum pump and other related equipment through the internet of things transmission protocol, and realizes real-time adjustment of the vacuum pump through remote monitoring.
By adopting the scheme, the beneficial effects obtained by the invention are as follows:
(1) Aiming at the problems that the optimal performance can be very complex and difficult to adjust in time in complex environments when the optimal performance is maintained under different working conditions by manually setting parameters and adjusting the parameters of the vacuum pump based on experience and expertise in the traditional vacuum pump control, the scheme uses a sensor to monitor the foundation environment condition, describes the foundation environment condition through a Markov decision process, learns the nonlinear mapping relation between the more complex foundation environment condition and a vacuum pump control strategy, and thus adjusts the working state of the vacuum pump more intelligently.
(2) Aiming at the influence of the foundation treatment effect when a common vacuum pump control strategy operates, including foundation vacuum pressure, sedimentation, groundwater level and pore water pressure, and in the whole foundation treatment process, the vacuum pump is always in a working state, even when the vacuum degree under a sealing film reaches a certain level, the vacuum pump still continuously works, the problem of wasting construction electricity is solved, the deep neural network is constructed to generate the control strategy, the neural network is trained by adopting a reinforcement learning algorithm, the strategy change updated each time is limited by introducing trust area constraint, the stability of the algorithm is improved, the excessive update of the control strategy is avoided, the flexibility of the vacuum pump control in the foundation treatment is increased, and the efficiency of the foundation treatment process is improved.
Drawings
FIG. 1 is a schematic flow chart of an AIoT-based intelligent linkage control method for a vacuum pump;
FIG. 2 is a schematic diagram of an AIoT-based intelligent linkage control system for a vacuum pump;
FIG. 3 is a flow chart of step S2
FIG. 4 is a flow chart of step S3;
the accompanying drawings are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate the invention and together with the embodiments of the invention, serve to explain the invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are only some, but not all embodiments of the invention; all other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
In the description of the present invention, it should be understood that the terms "upper," "lower," "front," "rear," "left," "right," "top," "bottom," "inner," "outer," and the like indicate orientation or positional relationships based on those shown in the drawings, merely to facilitate description of the invention and simplify the description, and do not indicate or imply that the devices or elements referred to must have a particular orientation, be constructed and operated in a particular orientation, and thus should not be construed as limiting the invention.
First embodiment, referring to fig. 1, the present invention provides an AIoT-based intelligent linkage control method for a vacuum pump, which includes the following steps:
step S1: the method comprises the steps of data acquisition and preprocessing, deploying a sensor to monitor foundation environmental conditions including vacuum pressure, sedimentation, groundwater level and pore water pressure, transmitting the foundation environmental conditions acquired by the sensor to a control center by using an Internet of things communication protocol, and receiving the foundation environmental conditions transmitted by the sensor by the control center and analyzing and preprocessing;
step S2: initializing a control strategy of the vacuum pump by using a Markov decision process, wherein the Markov decision process comprises a state space, an action space, a state transition probability, a reward function and a discount factor of the vacuum pump;
step S3: the control strategy is updated, a deep neural network for representing the control strategy is constructed, a state space of a vacuum pump is received as input, the probability of executing each action in a given state is output, the deep neural network is trained by using a TRPO algorithm, parameters of the deep neural network are iteratively optimized, and a final control strategy is generated;
step S4: and (3) remote monitoring and control, namely deploying a final control strategy to a control center, generating a control instruction, transmitting the control instruction to a vacuum pump and other related equipment through an Internet of things transmission protocol, and realizing real-time adjustment of the vacuum pump through remote monitoring.
In the second embodiment, referring to fig. 1 and 3, the control strategy initialization in step S2 specifically includes the following steps:
step S21: defining a control strategy, wherein the aim of defining the control strategy is to maximize the operation efficiency of the vacuum pump, minimize the energy consumption of the vacuum pump, and the constraint of the control strategy is the operation range of the vacuum pump;
step S22: establishing an MDP model, and modeling a control strategy of the vacuum pump through a Markov decision process, wherein the Markov decision process comprises a state space, an action space, a state transition probability, a reward function and a discount factor of the vacuum pump, and the following formula is adopted:
；
in the method, in the process of the invention,representing a Markov decision process,/->Representing state space, ++>Representing the action space->Representing state transition probabilities>Representing a reward function->Representing a discount factor;
step S23: defining MDP parameters, wherein the state space of the MDP comprises the running state of a vacuum pump, the vacuum level, sedimentation, the ground water level and pore water pressure, the action space of the MDP comprises the rotating speed of the vacuum pump, the opening degree of a valve and start-stop control, defining a reward function, setting a discount factor to be 0.7 by using a linear weighted summation method, and adopting the following formula:
；
in the method, in the process of the invention,rewarding the operating efficiency of the vacuum pump, +.>Rewarding the energy consumption of the vacuum pump, +.>Rewarding the operating range of the vacuum pump, +.>、/>、/>Weighting corresponding rewarding items;
step S24: defining a jackpot, maximizing the desired prize by selecting an optimal strategy, using the formula:
；
in the method, in the process of the invention,representing an optimal strategy->Representing policy ++>Is (are) desirable to be (are)>Representation->Status of moment->Representation->Action at time.
By executing the above operations, the parameters of the vacuum pump need to be adjusted based on experience and expertise by manually setting the parameters for the traditional vacuum pump control, so as to achieve the required vacuum level and the optimal treatment effect, the maintenance of the optimal performance becomes very complex when facing different working conditions, and the problem of difficult adjustment in time in the complex environment is solved.
Embodiment three, referring to fig. 1 and 4, based on the above embodiment, in step S3, the control policy update specifically includes the following steps:
step S31: the neural network design, construct the deep neural network used for representing the control strategy, the deep neural network is made up of input layer, hidden layer, output layer, the state space of MDP of input layer input, hidden layer uses tanh to activate the function to learn and represent complex functional relation from the input data, the formula used is as follows:
；
in the method, in the process of the invention,state information representing an input;
the output layer outputs the probability of performing the corresponding action in the given state, using the softmax function as the activation function of the output layer, using the following formula:
；
in the method, in the process of the invention,representing the +.>Element(s)>Representing the length of the input vector;
step S32: forward propagation, the forward propagation process of the deep neural network is constructed, and the following formula is used:
；
in the method, in the process of the invention,representing an activation value +.>Representing a weight matrix, +.>Representing the bias;
step S33: the objective function is designed to maximize the desired jackpot, using the following formula:
；
in the method, in the process of the invention,is a parameter of the policy, ++>Is the first timeTrack from start state to end state, +.>Is policy, & lt>Is at the moment +.>Rewards obtained at the department, ->Is a discount factor;
step S34: calculating a dominance estimate, estimating the relative superiority of each state-action pair using the following formula:
；
in the method, in the process of the invention,is a state-action pair->Is advantaged by->State-action pair->Is a function of the value of (c),is state->Is a function of the value of (2);
step S35: defining a loss function, and updating the direction of the parameters of the deep neural network by using a strategy gradient loss, wherein the following formula is used:
；
in the method, in the process of the invention,representing policy(s)>Representing +.>According to policy->Is subjected to a calculation of the expected value of the distribution of (c),representation strategy->Middle action->In state->The logarithmic probability of the lower case is related to the parameter->Is a gradient of (2);
step S36: selecting an optimization algorithm, and updating parameters of a strategy network by using a gradient ascent method to increase an objective function value, wherein the following formula is adopted:
；
in the method, in the process of the invention,is->Model parameters for the next iteration->Is the learning rate，/>Is the gradient of the loss function to the model parameters;
step S37: the trust zone is calculated, the KL divergence constraint control strategy is used for updating the trust zone of the amplitude, and the following formula is used:
；
in the method, in the process of the invention,indicating that the old policy is in the given state +.>The probability distribution of actions->Representing the probability distribution of actions of the current strategy in the same state,/-, for example>Representing +.>The integration is performed and the integration is performed,is a preset threshold value;
step S38: optimizing the objective function, and maximizing the objective function in the trust zone, wherein the formula is as follows:
；
in the method, in the process of the invention,is an optimized objective function;
step S39: and generating a control strategy, and adjusting parameters of the deep neural network through training optimization to generate a final control strategy.
By executing the operation, the control strategy is constructed by constructing the deep neural network to generate the control strategy, training the neural network by adopting the reinforcement learning algorithm, restricting the strategy change updated each time by introducing the trust zone constraint, improving the stability of the algorithm, avoiding excessive update of the control strategy, increasing the flexibility of the vacuum pump control in the foundation treatment and improving the efficiency of the foundation treatment process, and aiming at the influence of the foundation treatment effect when the common vacuum pump control strategy operates, including the foundation vacuum pressure, sedimentation, groundwater level and pore water pressure, and the vacuum pump is always in a working state in the whole foundation treatment process.
Fourth, referring to fig. 2, the embodiment is based on the above embodiment, and the AIoT-based intelligent linkage control system for a vacuum pump provided by the invention includes a data acquisition and preprocessing module, a control strategy initializing module, a control strategy updating module and a remote monitoring and control module;
the data acquisition and preprocessing module deploys a sensor to monitor foundation environmental conditions including vacuum pressure, sedimentation, groundwater level and pore water pressure, the foundation environmental conditions acquired by the sensor are transmitted to a control center by using an Internet of things communication protocol, the control center receives the foundation environmental conditions transmitted by the sensor and analyzes and preprocesses the foundation environmental conditions to obtain preprocessed data, and the preprocessed data is transmitted to the control strategy initialization module;
the control strategy initializing module receives data from the data acquisition and preprocessing module, defines a Markov decision process by using the received data, and sends the Markov decision process to the control strategy updating module;
the control strategy updating module receives a Markov decision process from the control strategy initializing module, constructs a deep neural network for representing the control strategy, receives the state of the vacuum pump as input, outputs the probability of executing each action in a given state, trains the deep neural network by using a TRPO algorithm, iteratively optimizes parameters of the deep neural network, generates a final control strategy, and sends the final control strategy to the remote monitoring and control module;
the remote monitoring and control module receives data from the control strategy updating module, deploys the final control strategy to the control center, generates control instructions, transmits the control instructions to the vacuum pump and other related equipment through the internet of things transmission protocol, and realizes real-time adjustment of the vacuum pump through remote monitoring.
It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus.
Although embodiments of the present invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made therein without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
The invention and its embodiments have been described above with no limitation, and the actual construction is not limited to the embodiments of the invention as shown in the drawings. In summary, if one of ordinary skill in the art is informed by this disclosure, a structural manner and an embodiment similar to the technical solution should not be creatively devised without departing from the gist of the present invention.
Claims (4)
1. The intelligent linkage control method of the vacuum pump based on AIoT is characterized by comprising the following steps of: the method comprises the following steps:
step S1: the method comprises the steps of data acquisition and preprocessing, deploying a sensor to monitor foundation environmental conditions including vacuum pressure, sedimentation, groundwater level and pore water pressure, transmitting the foundation environmental conditions acquired by the sensor to a control center by using an Internet of things communication protocol, and receiving the foundation environmental conditions transmitted by the sensor by the control center and analyzing and preprocessing;
step S2: initializing a control strategy of the vacuum pump by using a Markov decision process, wherein the Markov decision process comprises a state space, an action space, a state transition probability, a reward function and a discount factor of the vacuum pump;
step S3: the control strategy is updated, a deep neural network for representing the control strategy is constructed, a state space of a vacuum pump is received as input, the probability of executing each action in a given state is output, the deep neural network is trained by using a TRPO algorithm, parameters of the deep neural network are iteratively optimized, and a final control strategy is generated;
step S4: and (3) remote monitoring and control, namely deploying a final control strategy to a control center, generating a control instruction, transmitting the control instruction to a vacuum pump and other related equipment through an Internet of things transmission protocol, and realizing real-time adjustment of the vacuum pump through remote monitoring.
2. The AIoT-based intelligent coordinated control method of a vacuum pump of claim 1, wherein: in step S2, the control strategy is initialized, including the following steps:
step S21: defining a control strategy, wherein the aim of defining the control strategy is to maximize the operation efficiency of the vacuum pump, minimize the energy consumption of the vacuum pump, and the constraint of the control strategy is the operation range of the vacuum pump;
step S22: establishing an MDP model, and modeling a control strategy of the vacuum pump through a Markov decision process, wherein the Markov decision process comprises a state space, an action space, a state transition probability, a reward function and a discount factor of the vacuum pump, and the following formula is adopted:
；
in the method, in the process of the invention,representing a Markov decision process,/->Representing state space, ++>Representing the action space->Representing state transition probabilities>Representing a reward function->Representing a discount factor;
step S23: defining MDP parameters, wherein the state space of the MDP comprises the running state of a vacuum pump, the vacuum level, sedimentation, the ground water level and pore water pressure, the action space of the MDP comprises the rotating speed of the vacuum pump, the opening degree of a valve and start-stop control, defining a reward function, setting a discount factor to be 0.7 by using a linear weighted summation method, and adopting the following formula:
；
in the method, in the process of the invention,rewarding the operating efficiency of the vacuum pump, +.>Rewarding the energy consumption of the vacuum pump, +.>Rewarding the operating range of the vacuum pump, +.>、/>、/>Weighting corresponding rewarding items;
step S24: defining a jackpot, maximizing the desired prize by selecting an optimal strategy, using the formula:
；
in the method, in the process of the invention,representing an optimal strategy->Representing policy ++>Is (are) desirable to be (are)>Representation->Status of moment->Representation->Action at time.
3. The AIoT-based intelligent coordinated control method of a vacuum pump of claim 2, wherein: in step S3, the control strategy update includes the following steps:
step S31: the neural network design, construct the deep neural network used for representing the control strategy, the deep neural network is made up of input layer, hidden layer, output layer, the state space of MDP of input layer input, hidden layer uses tanh to activate the function to learn and represent complex functional relation from the input data, the formula used is as follows:
；
in the method, in the process of the invention,state information representing an input;
the output layer outputs the probability of performing the corresponding action in the given state, using the softmax function as the activation function of the output layer, using the following formula:
；
in the method, in the process of the invention,representing the +.>Element(s)>Representing the length of the input vector;
step S32: forward propagation, the forward propagation process of the deep neural network is constructed, and the following formula is used:
；
in the method, in the process of the invention,representing an activation value +.>Representing a weight matrix, +.>Representing the bias;
step S33: the objective function is designed to maximize the desired jackpot, using the following formula:
；
in the method, in the process of the invention,is a parameter of the policy, ++>Is a trace from an initial state to a termination state, +.>Is policy, & lt>Is at the moment +.>Rewards obtained at the department, ->Is a discount factor;
step S34: calculating a dominance estimate, estimating the relative superiority of each state-action pair using the following formula:
；
in the method, in the process of the invention,is a state-action pair->Is advantaged by->State-action pair->Value function of>Is state->Is a function of the value of (2);
step S35: defining a loss function, and updating the direction of the parameters of the deep neural network by using a strategy gradient loss, wherein the following formula is used:
；
in the method, in the process of the invention,representing policy(s)>Representing +.>According to policy->Is subjected to a calculation of the expected value of the distribution of (c),representation strategy->Middle action->In state->The logarithmic probability of the lower case is related to the parameter->Is a gradient of (2);
step S36: selecting an optimization algorithm, and updating parameters of a strategy network by using a gradient ascent method to increase an objective function value, wherein the following formula is adopted:
；
in the method, in the process of the invention,is->Model parameters for the next iteration->Is learning rate (I/O)>Is the gradient of the loss function to the model parameters;
step S37: the trust zone is calculated, the KL divergence constraint control strategy is used for updating the trust zone of the amplitude, and the following formula is used:
；
in the method, in the process of the invention,indicating that the old policy is in the given state +.>The probability distribution of actions->Representing the probability distribution of actions of the current strategy in the same state,/-, for example>Representing +.>Integrating->Is a preset threshold value;
step S38: optimizing the objective function, and maximizing the objective function in the trust zone, wherein the formula is as follows:
；
in the method, in the process of the invention,is an optimized objective function;
step S39: and generating a control strategy, and adjusting parameters of the deep neural network through training optimization to generate a final control strategy.
4. An AIoT-based intelligent linkage control system for a vacuum pump, configured to implement an AIoT-based intelligent linkage control method according to any one of claims 1 to 3, wherein: the system comprises a data acquisition and preprocessing module, a control strategy initialization module, a control strategy updating module and a remote monitoring and control module;
the data acquisition and preprocessing module deploys a sensor to monitor foundation environmental conditions including vacuum pressure, sedimentation, groundwater level and pore water pressure, the foundation environmental conditions acquired by the sensor are transmitted to a control center by using an Internet of things communication protocol, the control center receives the foundation environmental conditions transmitted by the sensor and analyzes and preprocesses the foundation environmental conditions to obtain preprocessed data, and the preprocessed data is transmitted to the control strategy initialization module;
the control strategy initializing module receives data from the data acquisition and preprocessing module, defines a Markov decision process by using the received data, and sends the Markov decision process to the control strategy updating module;
the control strategy updating module receives a Markov decision process from the control strategy initializing module, constructs a deep neural network for representing the control strategy, receives the state of the vacuum pump as input, outputs the probability of executing each action in a given state, trains the deep neural network by using a TRPO algorithm, iteratively optimizes parameters of the deep neural network, generates a final control strategy, and sends the final control strategy to the remote monitoring and control module;
the remote monitoring and control module receives data from the control strategy updating module, deploys the final control strategy to the control center, generates control instructions, transmits the control instructions to the vacuum pump and other related equipment through the internet of things transmission protocol, and realizes real-time adjustment of the vacuum pump through remote monitoring.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410075811.1A CN117588394B (en) | 2024-01-18 | 2024-01-18 | AIoT-based intelligent linkage control method and system for vacuum pump |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410075811.1A CN117588394B (en) | 2024-01-18 | 2024-01-18 | AIoT-based intelligent linkage control method and system for vacuum pump |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117588394A true CN117588394A (en) | 2024-02-23 |
| CN117588394B CN117588394B (en) | 2024-04-05 |
Family
ID=89910308
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410075811.1A Active CN117588394B (en) | 2024-01-18 | 2024-01-18 | AIoT-based intelligent linkage control method and system for vacuum pump |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117588394B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117780620A (en) * | 2024-02-28 | 2024-03-29 | 江苏凯威特斯半导体科技有限公司 | Vacuum pump performance test system and instrument based on internet of things |
| CN117780620B (en) * | 2024-02-28 | 2024-05-10 | 江苏凯威特斯半导体科技有限公司 | Vacuum pump performance test system and instrument based on internet of things |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150316282A1 (en) * | 2014-05-05 | 2015-11-05 | Board Of Regents, The University Of Texas System | Strategy for efficiently utilizing a heat-pump based hvac system with an auxiliary heating system |
| US20190266489A1 (en) * | 2017-10-12 | 2019-08-29 | Honda Motor Co., Ltd. | Interaction-aware decision making |
| CN111637614A (en) * | 2020-05-26 | 2020-09-08 | 内蒙古工业大学 | Intelligent control method for data center active ventilation floor |
| CN112241123A (en) * | 2020-10-23 | 2021-01-19 | 南京航空航天大学 | Aeroengine acceleration control method based on deep reinforcement learning |
| CN112329337A (en) * | 2020-10-23 | 2021-02-05 | 南京航空航天大学 | Aero-engine residual service life estimation method based on deep reinforcement learning |
| US20210367424A1 (en) * | 2020-05-19 | 2021-11-25 | Ruisheng Diao | Multi-Objective Real-time Power Flow Control Method Using Soft Actor-Critic |
| CN115526504A (en) * | 2022-09-30 | 2022-12-27 | 哈尔滨工业大学（深圳） | Energy-saving scheduling method and system for water supply system of pump station, electronic equipment and storage medium |
| CN115544899A (en) * | 2022-11-23 | 2022-12-30 | 南京邮电大学 | Water plant water intake pump station energy-saving scheduling method based on multi-agent deep reinforcement learning |
| WO2023136856A1 (en) * | 2022-01-17 | 2023-07-20 | Schlumberger Technology Corporation | Field pump equipment system |
| US20230239037A1 (en) * | 2022-01-24 | 2023-07-27 | Dongguan University Of Technology | Space-air-ground integrated uav-assisted iot data collectioncollection method based on aoi |
| CN116820134A (en) * | 2023-07-20 | 2023-09-29 | 智洋创新科技股份有限公司 | Unmanned aerial vehicle formation maintaining control method based on deep reinforcement learning |
| CN117037773A (en) * | 2023-08-22 | 2023-11-10 | 安徽省立医院(中国科学技术大学附属第一医院) | Insulin pump intelligent control system based on voice recognition |
-
2024
- 2024-01-18 CN CN202410075811.1A patent/CN117588394B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150316282A1 (en) * | 2014-05-05 | 2015-11-05 | Board Of Regents, The University Of Texas System | Strategy for efficiently utilizing a heat-pump based hvac system with an auxiliary heating system |
| US20190266489A1 (en) * | 2017-10-12 | 2019-08-29 | Honda Motor Co., Ltd. | Interaction-aware decision making |
| US20210367424A1 (en) * | 2020-05-19 | 2021-11-25 | Ruisheng Diao | Multi-Objective Real-time Power Flow Control Method Using Soft Actor-Critic |
| CN111637614A (en) * | 2020-05-26 | 2020-09-08 | 内蒙古工业大学 | Intelligent control method for data center active ventilation floor |
| CN112241123A (en) * | 2020-10-23 | 2021-01-19 | 南京航空航天大学 | Aeroengine acceleration control method based on deep reinforcement learning |
| CN112329337A (en) * | 2020-10-23 | 2021-02-05 | 南京航空航天大学 | Aero-engine residual service life estimation method based on deep reinforcement learning |
| WO2023136856A1 (en) * | 2022-01-17 | 2023-07-20 | Schlumberger Technology Corporation | Field pump equipment system |
| US20230239037A1 (en) * | 2022-01-24 | 2023-07-27 | Dongguan University Of Technology | Space-air-ground integrated uav-assisted iot data collectioncollection method based on aoi |
| CN115526504A (en) * | 2022-09-30 | 2022-12-27 | 哈尔滨工业大学（深圳） | Energy-saving scheduling method and system for water supply system of pump station, electronic equipment and storage medium |
| CN115544899A (en) * | 2022-11-23 | 2022-12-30 | 南京邮电大学 | Water plant water intake pump station energy-saving scheduling method based on multi-agent deep reinforcement learning |
| CN116820134A (en) * | 2023-07-20 | 2023-09-29 | 智洋创新科技股份有限公司 | Unmanned aerial vehicle formation maintaining control method based on deep reinforcement learning |
| CN117037773A (en) * | 2023-08-22 | 2023-11-10 | 安徽省立医院(中国科学技术大学附属第一医院) | Insulin pump intelligent control system based on voice recognition |
Non-Patent Citations (1)
| Title |
|---|
| 韩力群: "智能机器人技术丛书 机器智能与智能机器人", 31 March 2022, 北京：国防工业出版社, pages: 151 - 156 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117780620A (en) * | 2024-02-28 | 2024-03-29 | 江苏凯威特斯半导体科技有限公司 | Vacuum pump performance test system and instrument based on internet of things |
| CN117780620B (en) * | 2024-02-28 | 2024-05-10 | 江苏凯威特斯半导体科技有限公司 | Vacuum pump performance test system and instrument based on internet of things |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117588394B (en) | 2024-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107253195B (en) | A kind of carrying machine human arm manipulation ADAPTIVE MIXED study mapping intelligent control method and system | |
| CN103235620A (en) | Greenhouse environment intelligent control method based on global variable prediction model | |
| CN117588394B (en) | AIoT-based intelligent linkage control method and system for vacuum pump | |
| CN105825230A (en) | Soil moisture status prediction method and system based on DBN (Deep Belief Network) model | |
| CN111491358B (en) | Adaptive modulation and power control system based on energy acquisition and optimization method | |
| CN101763036A (en) | Lysine fermentation process feeding prediction control system and method based on fuzzy neural network | |
| CN101577003A (en) | Image segmenting method based on improvement of intersecting visual cortical model | |
| CN115983131B (en) | Aquatic product growth oxygen content regulation and control method and device | |
| CN105974801A (en) | Greenhouse environment modeling and control technology based on data drive | |
| CN113382060B (en) | Unmanned aerial vehicle track optimization method and system in Internet of things data collection | |
| CN105511270B (en) | A kind of PID controller parameter optimization method and system based on coevolution | |
| CN117077671A (en) | Interactive data generation method and system | |
| CN114967626A (en) | Multi-source federal environment control method for plant factory crop cultivation | |
| CN109711068A (en) | Elevation of building intelligentized design system | |
| CN115167419A (en) | Robot path planning method based on DQN algorithm | |
| Lyu et al. | Design of irrigation control system for vineyard based on lora wireless communication and dynamic neural network | |
| CN113110061A (en) | Intelligent irrigation fuzzy control method and system based on improved particle swarm optimization | |
| Fourati et al. | A greenhouse neural control using generalized and specialized learning | |
| Nkom et al. | Development of an Integrated Feedback and Feed-Forward Algorithm for Smart Irrigation Based on Fuzzy Logic | |
| KR100310609B1 (en) | Coefficient adjusting apparatus of proportional integral derivative(pid) controller | |
| CN117369286B (en) | Dynamic positioning control method for ocean platform | |
| CN117193025B (en) | Control method and device of dosing machine, electronic equipment and storage medium | |
| CN115138653B (en) | Intelligent flushing control method and system | |
| CN117784849A (en) | Automatic control system of refrigeration constant temperature tank based on artificial intelligence | |
| CN117252095A (en) | Micro-irrigation unit pipe network arrangement optimization design method based on reinforcement learning algorithm |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |