CN118153783A - Intelligent water-saving pipeline cleaning optimization system based on artificial intelligence - Google Patents
Intelligent water-saving pipeline cleaning optimization system based on artificial intelligence Download PDFInfo
- Publication number
- CN118153783A CN118153783A CN202410565984.1A CN202410565984A CN118153783A CN 118153783 A CN118153783 A CN 118153783A CN 202410565984 A CN202410565984 A CN 202410565984A CN 118153783 A CN118153783 A CN 118153783A
- Authority
- CN
- China
- Prior art keywords
- cleaning
- pipeline
- strategy
- water
- action
- 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
- 238000004140 cleaning Methods 0.000 title claims abstract description 204
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 11
- 238000005457 optimization Methods 0.000 title claims abstract description 11
- 230000009471 action Effects 0.000 claims abstract description 56
- 230000006870 function Effects 0.000 claims abstract description 50
- 238000000034 method Methods 0.000 claims abstract description 48
- 230000008859 change Effects 0.000 claims abstract description 35
- 230000000694 effects Effects 0.000 claims abstract description 27
- 238000005070 sampling Methods 0.000 claims abstract description 26
- 238000013528 artificial neural network Methods 0.000 claims abstract description 9
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 103
- 210000002569 neuron Anatomy 0.000 claims description 18
- 239000012459 cleaning agent Substances 0.000 claims description 15
- 238000001514 detection method Methods 0.000 claims description 15
- 230000001105 regulatory effect Effects 0.000 claims description 12
- 230000004913 activation Effects 0.000 claims description 8
- 230000000875 corresponding effect Effects 0.000 claims description 6
- 229910001385 heavy metal Inorganic materials 0.000 claims description 6
- 238000012544 monitoring process Methods 0.000 claims description 6
- 238000012549 training Methods 0.000 claims description 6
- 230000003044 adaptive effect Effects 0.000 claims description 5
- QVGXLLKOCUKJST-UHFFFAOYSA-N atomic oxygen Chemical compound [O] QVGXLLKOCUKJST-UHFFFAOYSA-N 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 230000003993 interaction Effects 0.000 claims description 3
- 229910052760 oxygen Inorganic materials 0.000 claims description 3
- 239000001301 oxygen Substances 0.000 claims description 3
- 230000008569 process Effects 0.000 abstract description 18
- 239000011159 matrix material Substances 0.000 description 8
- 239000002699 waste material Substances 0.000 description 7
- 230000007613 environmental effect Effects 0.000 description 4
- 238000009472 formulation Methods 0.000 description 4
- 239000000203 mixture Substances 0.000 description 4
- 230000008713 feedback mechanism Effects 0.000 description 3
- 230000003068 static effect Effects 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- JEIPFZHSYJVQDO-UHFFFAOYSA-N iron(III) oxide Inorganic materials O=[Fe]O[Fe]=O JEIPFZHSYJVQDO-UHFFFAOYSA-N 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000010865 sewage Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Cleaning In General (AREA)
Abstract
The invention relates to the technical field of pipeline cleaning, and particularly discloses an intelligent water-saving pipeline cleaning optimization system based on artificial intelligence. The scheme combines a depth deterministic strategy gradient algorithm and a self-adaptive sampling control method to formulate a cleaning strategy, so that the pipeline cleaner can learn autonomously and adapt to the change of a cleaning environment; the neural network is used for constructing a deterministic strategy function and an action value function, and the pipeline cleaner can intelligently adjust cleaning parameters through real-time rewarding feedback, so that efficient learning and optimizing processes are realized; the stability and reliability of the current state are evaluated in real time by calculating the change rate and the confidence of the state space, and the cleaning parameters and the sampling frequency of the sensor are dynamically adjusted according to the actual situation, so that the method is better suitable for the change of the environment, and the cleaning effect and the cleaning efficiency are improved.
Description
Technical Field
The invention relates to the technical field of pipeline cleaning, in particular to an intelligent water-saving pipeline cleaning optimization system based on artificial intelligence.
Background
The household pipeline cleaning is critical to maintaining the normal operation of a pipeline system, preventing blockage and prolonging the service life of a pipeline, and the traditional pipeline cleaning method is usually static, lacks intelligent decision, and has the problems of resource waste and high labor cost; the general pipeline cleaning method lacks a real-time feedback mechanism, cannot be timely adjusted and optimized according to the actual cleaning effect and environmental change, and has the problems that only the cleaning effect is focused, and factors such as cleaning cost, cleaning time and the like are ignored; the general strategy formulation method is often based on experience or rules, and cannot flexibly adjust the strategy according to real-time environmental changes, especially in a complex pipeline cleaning environment, a plurality of states and parameter changes are involved, and the problem that the complexity cannot be adapted exists.
Disclosure of Invention
Aiming at the problems of resource waste and high labor cost caused by the fact that the traditional pipeline cleaning method is usually static and lacks intelligent decision, the scheme adopts a depth deterministic strategy gradient algorithm and an adaptive sampling control method to combine and formulate a cleaning strategy, so that the pipeline cleaner can autonomously learn and adapt to the change of the cleaning environment, and unnecessary resource waste and labor cost are reduced; aiming at the problems that a general pipeline cleaning method lacks a real-time feedback mechanism, cannot be timely adjusted and optimized according to actual cleaning effects and environmental changes, only focuses on the cleaning effects and neglects factors such as cleaning cost and cleaning time, the scheme uses a neural network to construct a deterministic strategy function and an action value function, and the pipeline cleaner can intelligently adjust cleaning parameters through real-time rewarding feedback to realize efficient learning and optimizing processes and comprehensively consider a plurality of factors, so that the pipeline cleaner can better balance the cleaning effects, cost and time in the learning process; aiming at the problems that a general strategy formulation method is often based on experience or rules, the strategy cannot be flexibly adjusted according to real-time environment changes, and particularly in a complex pipeline cleaning environment, a plurality of states and parameter changes are involved, and the complexity cannot be adapted.
The intelligent water-saving pipeline cleaning optimization system based on artificial intelligence comprises a thermal circulation module, a water quality detection module and a pipeline cleaning module;
The thermal circulation module comprises a hot water source, water outlet points, water using points, a hot water pipe and a water return pipe, the water flow direction in the pipeline is set, hot water flows out from the water outlet points, is conveyed to each water using point along the hot water pipe, flows through the water return pipe and is conveyed back to the hot water source, thermal circulation information is formed, and the thermal circulation information is transmitted to the water quality detection module;
The water quality detection module receives thermal cycle information from the thermal cycle module, deploys a sensor at each water consumption point, constructs a cleaning strategy and sets a cleaning instruction; the sensor is used for monitoring water quality data, including pH value, turbidity, solubility and heavy metal amount, the cleaning strategy is constructed based on a depth deterministic strategy gradient algorithm and a self-adaptive sampling control method, the cleaning instruction consists of a cleaning switch and cleaning parameters, the cleaning parameters comprise a water temperature value, a water pressure value and a cleaning agent concentration, whether the cleaning effect of the pipeline reaches the standard is judged according to the monitored water quality data, and if the cleaning effect of the pipeline reaches the standard, the cleaning instruction is set to be zero; otherwise, setting a cleaning switch to be 1, setting a cleaning parameter according to a cleaning strategy, and sending a cleaning instruction to the pipeline cleaning module;
the pipeline cleaning module receives a cleaning instruction from the water quality detection module, a pipeline cleaner is installed, the longest cleaning time is set, the pipeline cleaner is used for setting the water temperature and the water pressure of pipeline cleaning, releasing a cleaning agent, analyzing the cleaning instruction, and stopping cleaning if a cleaning switch is 0; otherwise, continuing to clean according to the cleaning parameters until the longest cleaning time is reached.
Further, in the water quality detection module, the construction of the cleaning strategy includes the following steps:
Step S1: strategy initialization, namely initializing a cleaning strategy by using a depth deterministic strategy gradient algorithm, wherein the strategy comprises defining a state space and an action space, defining a reward function, initializing an action, and constructing a deterministic strategy function and an action value function;
Step S2: self-adaptive sampling control, namely calculating confidence coefficient according to the change rate of a state space, wherein the change rate of the state space is obtained by monitoring a sensor, and the cleaning parameters, the sampling frequency and the mode of the sensor are dynamically adjusted;
Step S3: strategy training, namely, iteratively training a cleaning strategy through interaction between a pipeline cleaner and a cleaning environment, and maximizing a pipeline cleaning effect;
Step S4: and evaluating performance of the cleaning strategy, feeding the performance of the cleaning strategy back to the cleaning strategy, and adjusting and optimizing the cleaning strategy according to actual requirements.
Further, in step S1, the policy initialization includes the following steps:
Step S11: defining a state space and an action space, wherein the state space is water quality data monitored by a sensor in real time, the state comprises pH value, turbidity, solubility of oxygen and heavy metal amount, the action space is a cleaning parameter regulated by a pipeline cleaner, and the actions comprise regulating a water temperature value, regulating a water pressure value and regulating a cleaning agent concentration;
step S12: defining a reward function, and designing the reward function as a feedback signal of the pipeline cleaner by combining the pipeline cleaning effect, the cleaning cost and the cleaning time, wherein the following formula is used:
；
In the method, in the process of the invention, Representing a bonus function that is based on the received data,The cleaning effect of the pipeline is shown,The cost of the cleaning is indicated and,The time for the cleaning is indicated and the time for the cleaning is indicated,Coefficients representing trade-offs of different factors;
Step S13: initializing an action, defining a value range of the action, and initializing the action to a random value in the value range;
Step S14: constructing a deterministic strategy function, namely constructing the deterministic strategy function by using a neural network, wherein the deterministic strategy function represents the optimal action taken by the pipeline cleaner in a given state, and an input layer comprises 4 neurons for inputting the state; the hidden layer is designed as a fully connected layer containing 64 neurons, connected using a ReLU activation function; the output layer is designed as a full-connection layer containing 3 neurons, and outputs the value of the corresponding action, and the formula is as follows:
；
In the method, in the process of the invention, The state of the display is indicated and,The action is represented by an action which,Representing the function of the ReLU activation,A matrix of weights representing the hidden layer is presented,The bias vector of the hidden layer is represented,Representing a weight matrix of the output layer,A bias vector representing an output layer;
step S15: constructing an action value function, namely constructing the action value function by using a neural network, wherein the action value represents the expected return of a specific action taken by the pipeline cleaner under a given state, and the input layer comprises 7 neurons for inputting a state-action pair; the hidden layer is designed as a fully connected layer containing 64 neurons, connected using a ReLU activation function; the output layer comprises a neuron and outputs a corresponding action value, and the formula is as follows:
；
In the method, in the process of the invention, The function of the action value is represented,A parameter representing a function of the action value,A state-action pair is represented and,A matrix of weights representing the hidden layer is presented,The bias vector of the hidden layer is represented,Representing a weight matrix of the output layer,Representing the offset vector of the output layer.
Further, in step S2, the adaptive sampling control includes the steps of:
Step S21: calculating the confidence coefficient, estimating the confidence coefficient of the current state based on the change rate of the state space, and if the change rate is smaller, the current state is relatively stable and the confidence coefficient is higher; conversely, if the rate of change is greater, then the confidence is lower, and the formula is as follows:
；
In the method, in the process of the invention, The degree of confidence is indicated and,The rate of change of a certain state in the state space,Indicating the maximum rate of change of this state,Standard deviation representing the feature;
Step S22: adjusting the cleaning parameters, dynamically adjusting the cleaning parameters according to the change rate and the confidence coefficient of the state space, and keeping the current cleaning parameters unchanged when the change rate is small and the confidence coefficient is high; when the change rate is large or the confidence is low, the water temperature, the water pressure and the concentration of the cleaning agent are adjusted;
Step S23: adjusting the sampling frequency and mode of the sensor according to the confidence coefficient, and reducing the sampling frequency to save energy and resources when the confidence coefficient is higher; when the confidence is lower, the sampling frequency is increased to obtain more data to evaluate the current state.
By adopting the scheme, the beneficial effects obtained by the invention are as follows:
(1) Aiming at the problems of resource waste and high labor cost caused by the lack of intelligent decision making in the conventional pipeline cleaning method, the scheme adopts a depth deterministic strategy gradient algorithm and a self-adaptive sampling control method to combine and formulate a cleaning strategy, so that the pipeline cleaner can autonomously learn and adapt to the change of the cleaning environment, and unnecessary resource waste and labor cost are reduced.
(2) Aiming at the problems that a general pipeline cleaning method lacks a real-time feedback mechanism, cannot be timely adjusted and optimized according to actual cleaning effects and environmental changes, only focuses on the cleaning effects and neglects factors such as cleaning cost and cleaning time, the scheme uses a neural network to construct a deterministic strategy function and an action value function, through real-time rewarding feedback, a pipeline cleaner can intelligently adjust cleaning parameters, an efficient learning and optimizing process is realized, and a plurality of factors are comprehensively considered, so that the pipeline cleaner can better balance the cleaning effects, cost and time in the learning process.
(3) Aiming at the problems that a general strategy formulation method is often based on experience or rules, the strategy cannot be flexibly adjusted according to real-time environment changes, and particularly in a complex pipeline cleaning environment, a plurality of states and parameter changes are involved, and the complexity cannot be adapted.
Drawings
FIG. 1 is a schematic diagram of an intelligent water-saving pipeline cleaning optimization system based on artificial intelligence;
Fig. 2 is a thermal cycle information chart of the pipe cleaning according to the present invention.
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.
Referring to fig. 1 and 2, the intelligent water-saving pipeline cleaning optimization system based on artificial intelligence provided by the invention comprises a thermal circulation module, a water quality detection module and a pipeline cleaning module;
The thermal circulation module comprises a hot water source, water outlet points, water using points, a hot water pipe and a water return pipe, the water flow direction in the pipeline is set, hot water flows out from the water outlet points, is conveyed to each water using point along the hot water pipe, flows through the water return pipe and is conveyed back to the hot water source, thermal circulation information is formed, and the thermal circulation information is transmitted to the water quality detection module;
The water quality detection module receives thermal cycle information from the thermal cycle module, deploys a sensor at each water consumption point, constructs a cleaning strategy and sets a cleaning instruction; the sensor is used for monitoring water quality data, including pH value, turbidity, solubility and heavy metal amount, the cleaning strategy is constructed based on a depth deterministic strategy gradient algorithm and a self-adaptive sampling control method, the cleaning instruction consists of a cleaning switch and cleaning parameters, the cleaning parameters comprise a water temperature value, a water pressure value and a cleaning agent concentration, whether the cleaning effect of the pipeline reaches the standard is judged according to the monitored water quality data, and if the cleaning effect of the pipeline reaches the standard, the cleaning instruction is set to be zero; otherwise, setting a cleaning switch to be 1, setting a cleaning parameter according to a cleaning strategy, and sending a cleaning instruction to the pipeline cleaning module;
the pipeline cleaning module receives a cleaning instruction from the water quality detection module, a pipeline cleaner is installed, the longest cleaning time is set, the pipeline cleaner is used for setting the water temperature and the water pressure of pipeline cleaning, releasing a cleaning agent, analyzing the cleaning instruction, and stopping cleaning if a cleaning switch is 0; otherwise, continuing to clean according to the cleaning parameters until the longest cleaning time is reached.
In a second embodiment, referring to fig. 1, the water quality detection module according to the above embodiment, the cleaning strategy is constructed, and the method includes the following steps:
Step S1: strategy initialization, namely initializing a cleaning strategy by using a depth deterministic strategy gradient algorithm, wherein the strategy comprises defining a state space and an action space, defining a reward function, initializing an action, and constructing a deterministic strategy function and an action value function;
Step S2: self-adaptive sampling control, namely calculating confidence coefficient according to the change rate of a state space, wherein the change rate of the state space is obtained by monitoring a sensor, and the cleaning parameters, the sampling frequency and the mode of the sensor are dynamically adjusted;
Step S3: strategy training, namely, iteratively training a cleaning strategy through interaction between a pipeline cleaner and a cleaning environment, and maximizing a pipeline cleaning effect;
Step S4: and evaluating performance of the cleaning strategy, feeding the performance of the cleaning strategy back to the cleaning strategy, and adjusting and optimizing the cleaning strategy according to actual requirements.
By executing the operation, aiming at the problems of resource waste and high labor cost caused by the lack of intelligent decision in the conventional pipeline cleaning method, the scheme adopts a depth deterministic strategy gradient algorithm and an adaptive sampling control method to combine and formulate a cleaning strategy, so that the pipeline cleaner can autonomously learn and adapt to the change of the cleaning environment, and unnecessary resource waste and labor cost are reduced.
Embodiment three, referring to fig. 1, the embodiment is based on the above embodiment, and in step S1, the policy initialization includes the following steps:
Step S11: defining a state space and an action space, wherein the state space is water quality data monitored by a sensor in real time, the state comprises pH value, turbidity, solubility of oxygen and heavy metal amount, the action space is a cleaning parameter regulated by a pipeline cleaner, and the actions comprise regulating a water temperature value, regulating a water pressure value and regulating a cleaning agent concentration;
step S12: defining a reward function, and designing the reward function as a feedback signal of the pipeline cleaner by combining the pipeline cleaning effect, the cleaning cost and the cleaning time, wherein the following formula is used:
；
In the method, in the process of the invention, Representing a bonus function that is based on the received data,The cleaning effect of the pipeline is shown,The cost of the cleaning is indicated and,The time for the cleaning is indicated and the time for the cleaning is indicated,Coefficients representing trade-offs of different factors;
Step S13: initializing an action, defining a value range of the action, and initializing the action to a random value in the value range;
Step S14: constructing a deterministic strategy function, namely constructing the deterministic strategy function by using a neural network, wherein the deterministic strategy function represents the optimal action taken by the pipeline cleaner in a given state, and an input layer comprises 4 neurons for inputting the state; the hidden layer is designed as a fully connected layer containing 64 neurons, connected using a ReLU activation function; the output layer is designed as a full-connection layer containing 3 neurons, and outputs the value of the corresponding action, and the formula is as follows:
；
In the method, in the process of the invention, The state of the display is indicated and,The action is represented by an action which,Representing the function of the ReLU activation,A matrix of weights representing the hidden layer is presented,The bias vector of the hidden layer is represented,Representing a weight matrix of the output layer,A bias vector representing an output layer;
step S15: constructing an action value function, namely constructing the action value function by using a neural network, wherein the action value represents the expected return of a specific action taken by the pipeline cleaner under a given state, and the input layer comprises 7 neurons for inputting a state-action pair; the hidden layer is designed as a fully connected layer containing 64 neurons, connected using a ReLU activation function; the output layer comprises a neuron and outputs a corresponding action value, and the formula is as follows:
；
In the method, in the process of the invention, The function of the action value is represented,A parameter representing a function of the action value,A state-action pair is represented and,A matrix of weights representing the hidden layer is presented,The bias vector of the hidden layer is represented,Representing a weight matrix of the output layer,Representing the offset vector of the output layer.
By executing the operation, the pipeline cleaner can intelligently adjust the cleaning parameters through real-time reward feedback, realize efficient learning and optimizing processes, comprehensively consider a plurality of factors, and enable the pipeline cleaner to better balance the cleaning effect, cost and time in the learning process.
Fourth embodiment, referring to fig. 1, the adaptive sampling control in step S2, which is based on the above embodiment, includes the following steps:
Step S21: calculating the confidence coefficient, estimating the confidence coefficient of the current state based on the change rate of the state space, and if the change rate is smaller, the current state is relatively stable and the confidence coefficient is higher; conversely, if the rate of change is greater, then the confidence is lower, and the formula is as follows:
；
In the method, in the process of the invention, The degree of confidence is indicated and,The rate of change of a certain state in the state space,Indicating the maximum rate of change of this state,Standard deviation representing the feature;
Step S22: adjusting the cleaning parameters, dynamically adjusting the cleaning parameters according to the change rate and the confidence coefficient of the state space, and keeping the current cleaning parameters unchanged when the change rate is small and the confidence coefficient is high; when the change rate is large or the confidence is low, the water temperature, the water pressure and the concentration of the cleaning agent are adjusted;
Step S23: adjusting the sampling frequency and mode of the sensor according to the confidence coefficient, and reducing the sampling frequency to save energy and resources when the confidence coefficient is higher; when the confidence is lower, the sampling frequency is increased to obtain more data to evaluate the current state.
By executing the operation, the general strategy formulation method is often based on experience or rules, the strategy cannot be flexibly adjusted according to real-time environment changes, and particularly in a complex pipeline cleaning environment, a plurality of states and parameter changes are involved, and the problem that the complexity cannot be adapted exists.
In the fifth embodiment, referring to fig. 1, the embodiment is based on the above embodiment, in the pipe cleaning module, the longest cleaning time is 1 hour, and if the pipe cleaning effect does not reach the standard after the longest cleaning time is reached, the pipe maintenance is required.
Embodiment six, referring to fig. 1, the embodiment is based on the above embodiment, and in step S12, the cleaning cost includes water consumption, cleaning agent consumption, and power consumption.
An embodiment seven, referring to fig. 1, based on the above embodiment, in step S4, the performance evaluation includes collecting time, manpower, water resources and cleaning agent costs consumed in the cleaning process, comparing these costs with the cleaning effect, and evaluating the economical efficiency and efficiency of the cleaning strategy; collecting and analyzing historical cleaning data, and evaluating long-term effects and trends of cleaning strategies so as to optimize and improve; and evaluating whether the cleaned water flow is smoother, the water quality is improved, and dirt residues exist in the pipeline.
In the embodiment, the flow of pipeline cleaning is specifically that hot water is released by a hot water source, hot water flows out from a water outlet point, a cleaning agent is dissolved into a pipeline after passing through a pipeline cleaner, rust and dirt in the pipeline begin to be cleaned, water quality data is monitored by a sensor before the water flows to each water using point, cleaned sewage flows out from the water using point, the water flows circulate, and cleaning is stopped after the cleaning effect of the pipeline reaches the standard or the longest cleaning time is reached.
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. Intelligent water-saving pipeline cleaning optimization system based on artificial intelligence, which is characterized in that: the device comprises a thermal circulation module, a water quality detection module and a pipeline cleaning module;
The thermal circulation module comprises a hot water source, water outlet points, water using points, a hot water pipe and a water return pipe, the water flow direction in the pipeline is set, hot water flows out from the water outlet points, is conveyed to each water using point along the hot water pipe, flows through the water return pipe and is conveyed back to the hot water source, thermal circulation information is formed, and the thermal circulation information is transmitted to the water quality detection module;
The water quality detection module receives thermal cycle information from the thermal cycle module, deploys a sensor at each water consumption point, constructs a cleaning strategy and sets a cleaning instruction; the sensor is used for monitoring water quality data, including pH value, turbidity, solubility and heavy metal amount, the cleaning strategy is constructed based on a depth deterministic strategy gradient algorithm and a self-adaptive sampling control method, the cleaning instruction consists of a cleaning switch and cleaning parameters, the cleaning parameters comprise a water temperature value, a water pressure value and a cleaning agent concentration, whether the cleaning effect of the pipeline reaches the standard is judged according to the monitored water quality data, and if the cleaning effect of the pipeline reaches the standard, the cleaning instruction is set to be zero; otherwise, setting a cleaning switch to be 1, setting a cleaning parameter according to a cleaning strategy, and sending a cleaning instruction to the pipeline cleaning module;
the pipeline cleaning module receives a cleaning instruction from the water quality detection module, a pipeline cleaner is installed, the longest cleaning time is set, the pipeline cleaner is used for setting the water temperature and the water pressure of pipeline cleaning, releasing a cleaning agent, analyzing the cleaning instruction, and stopping cleaning if a cleaning switch is 0; otherwise, continuing to clean according to the cleaning parameters until the longest cleaning time is reached.
2. The intelligent water saving pipe cleaning optimization system based on artificial intelligence according to claim 1, wherein: in the water quality detection module, the construction of the cleaning strategy comprises the following steps:
Step S1: strategy initialization, namely initializing a cleaning strategy by using a depth deterministic strategy gradient algorithm, wherein the strategy comprises defining a state space and an action space, defining a reward function, initializing an action, and constructing a deterministic strategy function and an action value function;
Step S2: self-adaptive sampling control, namely calculating confidence coefficient according to the change rate of a state space, wherein the change rate of the state space is obtained by monitoring a sensor, and the cleaning parameters, the sampling frequency and the mode of the sensor are dynamically adjusted;
Step S3: strategy training, namely, iteratively training a cleaning strategy through interaction between a pipeline cleaner and a cleaning environment, and maximizing a pipeline cleaning effect;
Step S4: and evaluating performance of the cleaning strategy, feeding the performance of the cleaning strategy back to the cleaning strategy, and adjusting and optimizing the cleaning strategy according to actual requirements.
3. The intelligent water saving pipe cleaning optimization system based on artificial intelligence according to claim 2, wherein: in step S1, the policy initialization includes the following steps:
Step S11: defining a state space and an action space, wherein the state space is water quality data monitored by a sensor in real time, the state comprises pH value, turbidity, solubility of oxygen and heavy metal amount, the action space is a cleaning parameter regulated by a pipeline cleaner, and the actions comprise regulating a water temperature value, regulating a water pressure value and regulating a cleaning agent concentration;
step S12: defining a reward function, and designing the reward function as a feedback signal of the pipeline cleaner by combining the pipeline cleaning effect, the cleaning cost and the cleaning time;
Step S13: initializing an action, defining a value range of the action, and initializing the action to a random value in the value range;
Step S14: constructing a deterministic strategy function, namely constructing the deterministic strategy function by using a neural network, wherein the deterministic strategy function represents the optimal action taken by the pipeline cleaner in a given state, and an input layer comprises 4 neurons for inputting the state; the hidden layer is designed as a fully connected layer containing 64 neurons, connected using a ReLU activation function; the output layer is designed as a full-connection layer containing 3 neurons and outputs the value of the corresponding action;
Step S15: constructing an action value function, namely constructing the action value function by using a neural network, wherein the action value represents the expected return of a specific action taken by the pipeline cleaner under a given state, and the input layer comprises 7 neurons for inputting a state-action pair; the hidden layer is designed as a fully connected layer containing 64 neurons, connected using a ReLU activation function; the output layer comprises a neuron and outputs a corresponding action value.
4. An artificial intelligence based intelligent water saving pipeline cleaning optimization system according to claim 3, characterized in that: in step S2, the adaptive sampling control includes the steps of:
step S21: calculating the confidence coefficient, estimating the confidence coefficient of the current state based on the change rate of the state space, and if the change rate is smaller, the current state is relatively stable and the confidence coefficient is higher; conversely, if the rate of change is greater, the confidence is lower;
Step S22: adjusting the cleaning parameters, dynamically adjusting the cleaning parameters according to the change rate and the confidence coefficient of the state space, and keeping the current cleaning parameters unchanged when the change rate is small and the confidence coefficient is high; when the change rate is large or the confidence is low, the water temperature, the water pressure and the concentration of the cleaning agent are adjusted;
Step S23: adjusting the sampling frequency and mode of the sensor according to the confidence coefficient, and reducing the sampling frequency to save energy and resources when the confidence coefficient is higher; when the confidence is lower, the sampling frequency is increased to obtain more data to evaluate the current state.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410565984.1A CN118153783B (en) | 2024-05-09 | Intelligent water-saving pipeline cleaning optimization system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410565984.1A CN118153783B (en) | 2024-05-09 | Intelligent water-saving pipeline cleaning optimization system based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118153783A true CN118153783A (en) | 2024-06-07 |
| CN118153783B CN118153783B (en) | 2024-07-26 |
Family
ID=
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112327821A (en) * | 2020-07-08 | 2021-02-05 | 东莞市均谊视觉科技有限公司 | Intelligent cleaning robot path planning method based on deep reinforcement learning |
| CN113503885A (en) * | 2021-04-30 | 2021-10-15 | 山东师范大学 | Robot path navigation method and system based on sampling optimization DDPG algorithm |
| CN115672898A (en) * | 2022-12-30 | 2023-02-03 | 山东天弘化学有限公司 | Intelligent cleaning device and cleaning method for petrochemical engineering pipeline |
| KR20230128191A (en) * | 2022-02-25 | 2023-09-04 | 경희대학교 산학협력단 | An auto-adaptive controller tuning system for wastewater treatment process using data-driven smart decisions of offline reinforcement learning |
| CN117195443A (en) * | 2023-07-25 | 2023-12-08 | 苏州科技大学 | Rainwater pipe network optimization method based on depth deterministic strategy gradient algorithm |
| CN117434844A (en) * | 2023-12-08 | 2024-01-23 | 华南理工大学 | Papermaking sewage treatment optimization control method based on reinforcement learning |
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112327821A (en) * | 2020-07-08 | 2021-02-05 | 东莞市均谊视觉科技有限公司 | Intelligent cleaning robot path planning method based on deep reinforcement learning |
| CN113503885A (en) * | 2021-04-30 | 2021-10-15 | 山东师范大学 | Robot path navigation method and system based on sampling optimization DDPG algorithm |
| KR20230128191A (en) * | 2022-02-25 | 2023-09-04 | 경희대학교 산학협력단 | An auto-adaptive controller tuning system for wastewater treatment process using data-driven smart decisions of offline reinforcement learning |
| CN115672898A (en) * | 2022-12-30 | 2023-02-03 | 山东天弘化学有限公司 | Intelligent cleaning device and cleaning method for petrochemical engineering pipeline |
| CN117195443A (en) * | 2023-07-25 | 2023-12-08 | 苏州科技大学 | Rainwater pipe network optimization method based on depth deterministic strategy gradient algorithm |
| CN117434844A (en) * | 2023-12-08 | 2024-01-23 | 华南理工大学 | Papermaking sewage treatment optimization control method based on reinforcement learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9747544B2 (en) | Method and system for wastewater treatment based on dissolved oxygen control by fuzzy neural network | |
| Cheng et al. | Forecasting of wastewater treatment plant key features using deep learning-based models: A case study | |
| CN101576734B (en) | Dissolved oxygen control method based on dynamic radial basis function neural network | |
| KR20030097425A (en) | System and method for AI controlling waste-water treatment by neural network and back-propagation algorithm | |
| CN108536106B (en) | Aeration system dissolved oxygen online regulation and control method based on Kalman filtering-extreme learning machine | |
| Lee et al. | Universal workflow of artificial intelligence for energy saving | |
| Qiu et al. | Chilled water temperature resetting using model-free reinforcement learning: Engineering application | |
| Shioya et al. | Knowledge-based design and operation of bioprocess systems | |
| CN118153783B (en) | Intelligent water-saving pipeline cleaning optimization system based on artificial intelligence | |
| Ibrahim et al. | Smart irrigation system using a fuzzy logic method | |
| CN114671523A (en) | Sewage treatment energy-saving method based on predictive control | |
| CN118153783A (en) | Intelligent water-saving pipeline cleaning optimization system based on artificial intelligence | |
| He et al. | Prediction of electricity demand of China based on the analysis of decoupling and driving force | |
| Chen et al. | The predictive management in campus heating system based on deep reinforcement learning and probabilistic heat demands forecasting | |
| Kim et al. | Lstm filter for smart agriculture | |
| KR100935364B1 (en) | Unmanned remote total managing system for water treatment using fuzzy model according to change of water level and method thereof | |
| Liu et al. | Dynamic multi-objective optimization control for wastewater treatment process based on modal decomposition and hybrid neural network | |
| Chen et al. | Application of fuzzy logic to reduce modelling errors in PIDSP for FOPDT process control | |
| Gupta et al. | Multi-objective reinforcement learning based approach for user-centric power optimization in smart home environments | |
| Agustian et al. | NFT Hydroponic Control Using Mamdani Fuzzy Inference System | |
| Silva et al. | Framework for the development of a digital twin for solar water heating systems | |
| Argyropoulos | Soft sensor development and process control of anaerobic digestion | |
| Ateunkeng et al. | Hybrid supervised hierarchical control of a biological wastewater treatment plant | |
| Basavaraddi et al. | Optimizing Water Distribution Pump Scheduling with SVM-Based Machine Learning Using IoT Sensor Data | |
| Hernandez-del-Olmo et al. | Reinforcement learning techniques for the control of wastewater treatment plants |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant |