Intelligent water supply control method and system

Abstract

The invention relates to the technical field of intelligent water supply, and discloses an intelligent water supply control method and system, and the method comprises the steps: collecting the environment parameters of an urban water supply system at the current time; according to a preset fuzzy control model, the environmental parameters of the urban water supply system at the current time are calculated, and the water supply adjusting amount at the current time is obtained; inputting the environmental parameters of the urban water supply system at the current time and the water supply adjusting quantity into a preset LSTM neural network; outputting a water supply strategy of the water pump of the urban water supply system at the current time t through the LSTM neural network; and according to the water supply strategy of the water pump, controlling the valve opening of the water pump to enable the water pump to supply water in the urban water supply system. By adopting the fuzzy control model and the LSTM neural network, automatic water supply control is realized, the requirement of manual intervention is reduced, the workload of operators is reduced, and the construction efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent water supply technology, and more specifically, to an intelligent water supply control method and system. Background Art

[0002] With the acceleration of urbanization, urban water supply systems are facing increasingly severe challenges. On the one hand, the increase in urban population density has led to an increase in demand for water resources; on the other hand, external factors such as climate change have also brought uncertainty to the stability of urban water supply systems. In order to meet these challenges, traditional water supply system management methods can no longer meet the needs of modern cities, and more intelligent technical means need to be introduced to improve water supply efficiency and service quality.

[0003] Traditional urban water supply systems mainly rely on manual monitoring and manual adjustment, which not only consumes a lot of manpower and material resources, but also often cannot respond to emergencies in the water network in a timely manner due to delayed information feedback. In addition, when faced with seasonal water consumption peaks or extreme weather events, traditional water supply systems lack sufficient flexibility to make dynamic adjustments.

[0004] Therefore, it is particularly important to develop a new intelligent water supply control method so that the water supply system can realize automatic water supply according to environmental parameters, improve water supply stability and reduce water supply accident rate. Summary of the invention

[0005] In order to solve the above-mentioned technical problems, the present application is proposed to provide an intelligent water supply control method and system, which can realize automatic water supply according to environmental parameters, improve water supply stability, and reduce water supply accident rate.

[0006] In a first aspect, the present invention provides an intelligent water supply control method, comprising: collecting environmental parameters of a city water supply system at a current time, including the temperature, humidity, and water supply pressure of the city water supply system; calculating the environmental parameters of the city water supply system at the current time according to a preset fuzzy control model to obtain a water supply regulation amount at the current time; inputting the environmental parameters of the city water supply system at the current time and the water supply regulation amount into a preset LSTM neural network, wherein the LSTM neural network has a multi-layer neuron structure and is trained based on the environmental parameters of the city water supply system at a historical time, the water supply regulation amount, and a target water supply strategy through a water pump, wherein for any weight w of the associated two adjacent layers of neurons in the LSTM neural network: and Where w(n), w(n+1), and w(n-1) are the values ​​of weight w after the nth, n+1th, and n-1th trainings, respectively; ρ is the preset learning rate; and E(n) is the value of the error function corresponding to the LSTM neural network after the nth training. is the partial differential symbol, α is the preset momentum term coefficient, L(n) is the value of the loss function after the nth training of the LSTM neural network, and β is the preset smoothing coefficient; the water supply strategy of the water pump of the urban water supply system at the current time t is output through the LSTM neural network, including the flow rate and head of the water pump; according to the water supply strategy of the water pump, the valve opening of the water pump is controlled to make the water pump supply water to supply water in the urban water supply system.

[0007] Optionally, the aforementioned intelligent water supply control method, before calculating the environmental parameters of the urban water supply system at the current time according to the preset fuzzy control model, further comprises: setting a minimum value P of the water supply pressure of the urban water supply system in the fuzzy control model. min and the maximum value P max , to determine the fuzzy domain of the water supply pressure of the urban water supply system; calculate the minimum value P of the water supply pressure of the urban water supply system min and the maximum value P max The mean value P mid ; Set a membership function for the water supply pressure of the urban water supply system Wherein, P is a variable representing the water supply pressure of the urban water supply system.

[0008] Optionally, the aforementioned intelligent water supply control method, before calculating the environmental parameters of the urban water supply system at the current time according to the preset fuzzy control model, further comprises: detecting the geographical location of the urban water supply system to determine the area where the urban water supply system is located; querying the historical maximum humidity H of the area where the urban water supply system is located; hmax and the lowest humidity in history hmin ; Take the minimum value of humidity of the city water supply system at the current time Take the maximum value of the humidity of the city water supply system at the current time To determine the fuzzy domain of humidity of the city water system at the current time, where P bmin is the preset minimum basic humidity value, H bmax is the preset maximum basic humidity value, and ΔH is the preset humidity adjustment range.

[0009] Optionally, the aforementioned intelligent water supply control method further includes: detecting the weather conditions of the area where the urban water supply system is located at the current time; setting a weather influence factor γ according to the weather conditions of the area where the urban water supply system is located at the current time. wea ; According to the weather influencing factor, the minimum and maximum values ​​of the humidity of the urban water supply system at the current time are adjusted, and the minimum value of the humidity of the urban water supply system at the current time is The maximum humidity of the city water supply system at the current time

[0010] Optionally, the aforementioned intelligent water supply control method, before calculating the environmental parameters of the urban water supply system at the current time according to the preset fuzzy control model, further includes: recording multiple fuzzy rules in the fuzzy control model in a preset file, wherein the file is used to query required fuzzy rules when using the fuzzy control model for calculation; sorting the multiple fuzzy calculation rules according to the number of times the multiple fuzzy rules are used recorded in advance; and adjusting the positions of the multiple fuzzy rules in the file according to the sorting of the multiple fuzzy calculation rules.

[0011] Optionally, the aforementioned intelligent water supply control method, before inputting the environmental parameters of the urban water supply system at the current time and the water supply regulation amount into the preset LSTM neural network, further comprises: when training the LSTM neural network, setting an unbalanced loss L1=w for the LSTM neural network e (y1-y2) 2 , where y1 is the water supply strategy output during the LSTM neural network training, y2 is the target water supply strategy that should be output during the LSTM neural network training, and w e Indicates the impact level of the gap between the water supply strategy output by the LSTM neural network and the target water supply strategy that should be output on the occurrence rate of water supply accidents; sets a regularization loss for the LSTM neural network Wherein, τ is the preset regularization coefficient, z is the total number of weights in the LSTM neural network, and w i is the i-th weight; sets the loss function L=L1+L2 for the LSTM neural network; and calculates the value of the loss function L(n) after the n-th training of the LSTM neural network.

[0012] Optionally, the aforementioned intelligent water supply control method, before inputting the environmental parameters of the urban water supply system at the current time and the water supply regulation amount into the preset LSTM neural network, further comprises: when the LSTM neural network is trained for the nth time, calculating the current statistical period And round its value, where t is a positive integer; compare the current statistical period r with the recorded historical statistical period r h The size of the current statistical period r is greater than the historical statistical period r h When the loss function L(n) of the LSTM neural network after the nth training is compared with the pre-recorded historical loss function L h The value of the loss function L(n) after the nth training of the LSTM neural network is less than the historical loss function L h When the value of is, the learning rate ρ is reduced; the historical statistical period r is updated using the current statistical period r h , use the value of the loss function L(n) after the nth training of the LSTM neural network to update the historical loss function L h .

[0013] Optionally, in the aforementioned intelligent water supply control method, the step of reducing the learning rate ρ comprises: calculating a target learning rate ρ tar =ρ0D 1+r , where D is between 0 and 1; determine the target learning rate ρ tar is lower than the preset minimum learning rate; if the result is no, the target learning rate ρ tar When the judgment result is yes, the value of the learning rate ρ is updated according to the minimum learning rate.

[0014] In a second aspect, the present invention provides an intelligent water supply control system, comprising: an environmental parameter collection module, which collects the environmental parameters of the urban water supply system at the current time, including the temperature, humidity, and water supply pressure of the urban water supply system; a water supply quantity calculation module, which calculates the environmental parameters of the urban water supply system at the current time according to a preset fuzzy control model to obtain the water supply regulation at the current time; a water supply strategy calculation module, which inputs the environmental parameters of the urban water supply system at the current time and the water supply regulation into a preset LSTM neural network, wherein the LSTM neural network has a multi-layer neuron structure and is trained based on the environmental parameters of the urban water supply system at historical time, the water supply regulation, and the target water supply strategy through the water pump, wherein, for any weight w of the associated two adjacent layers of neurons in the LSTM neural network: and Where w(n), w(n+1), and w(n-1) are the values ​​of weight w after the nth, n+1th, and n-1th trainings, respectively; ρ is the preset learning rate; and E(n) is the value of the error function corresponding to the LSTM neural network after the nth training. is the partial differential symbol, α is the preset momentum term coefficient, L(n) is the value of the loss function after the nth training of the LSTM neural network, and β is the preset smoothing coefficient; the water supply strategy of the water pump of the urban water supply system at the current time t is output through the LSTM neural network, including the flow rate and head of the water pump; the water supply module controls the valve opening of the water pump according to the water supply strategy of the water pump so that the water pump supplies water to supply water in the urban water supply system.

[0015] The above one or more technical solutions of the present invention have at least one or more of the following beneficial effects:

[0016] According to the technical solution of the present invention, by adopting a preset fuzzy control model and LSTM neural network, automatic water supply control is realized, the need for manual intervention is reduced, and the workload of operators is reduced. By real-time monitoring and analysis of environmental parameters of the urban water supply system, the water supply strategy can be adjusted in time to ensure that the water supply stability is in an ideal state during the water supply process. In the overall process, the calculation efficiency is improved by the fuzzy control method, and the calculation accuracy is improved by the LSTM neural network, so the technical solution of the present invention can take into account both efficiency and accuracy when realizing water supply control. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0018] Figure 1 is a flow chart of a smart water supply control method according to an embodiment of the present application;

[0019] Figure 2 is a partial flow chart of the intelligent water supply control method according to an embodiment of the present application;

[0020] Figure 3 is another layout flow chart of the intelligent water supply control method according to an embodiment of the present application;

[0021] Figure 4 is another partial flow chart of the intelligent water supply control method according to an embodiment of the present application;

[0022] Figure 5 A partial schematic diagram of a smart water supply control method according to an embodiment of the present application;

[0023] Figure 6 4 is a block diagram of an intelligent water supply control system according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] Some embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0025] like Figure 1 As shown, in one embodiment of the present invention, a smart water supply control method is provided, comprising:

[0026] Step S110, collecting environmental parameters of the city water supply system at the current time, including temperature, humidity, and water pressure of the city water supply system.

[0027] Step S120, calculating the environmental parameters of the city water supply system at the current time according to the preset fuzzy control model to obtain the water supply regulation amount at the current time.

[0028] In this embodiment, the fuzzy control model can handle the uncertainty and nonlinear problems in the water supply regulation amount estimation process, and the fuzzy control calculation process is relatively fast.

[0029] Step S130, inputting the environmental parameters and water supply regulation of the city water supply system at the current time into a preset LSTM neural network, the LSTM neural network has a multi-layer neuron structure, and is trained based on the environmental parameters, water supply regulation of the city water supply system at historical time and the target water supply strategy through the water pump, wherein, for any weight w of the associated two adjacent layers of neurons in the LSTM neural network: and Wherein w(n), w(n+1), and w(n-1) are the values ​​of the weight w after the nth, n+1th, and n-1th trainings, respectively; ρ is a preset learning rate; E(n) is the value of the error function corresponding to the LSTM neural network after the nth training; θ is the partial differential symbol; α is the preset momentum term coefficient; L(n) is the value of the loss function of the LSTM neural network after the nth training; and β is a preset smoothing coefficient.

[0030] In this embodiment, the LSTM neural network can predict future water supply demand based on current environmental parameters by learning historical data, thereby adjusting the water supply strategy more accurately to ensure that the water supply volume matches the water supply pressure. The training of the LSTM neural network adopts an efficient update formula, and by introducing momentum terms and adaptive learning rate adjustment strategies, the convergence speed and computational efficiency of the model are improved, making the calculation of the water supply strategy faster.

[0031] Step S140, outputting the water supply strategy of the water pump of the urban water supply system at the current time t through the LSTM neural network, including the flow rate and lift of the water pump.

[0032] Step S150, according to the water supply strategy of the water pump, controlling the valve opening of the water pump to enable the water pump to supply water to the city water supply system.

[0033] According to the technical solution of this embodiment, by adopting a preset fuzzy control model and LSTM neural network, automatic water supply control is realized, the need for manual intervention is reduced, and the workload of operators is reduced. By real-time monitoring and analysis of environmental parameters of the urban water supply system, the water supply strategy can be adjusted in time to ensure that the water supply stability is in an ideal state during the water supply process. In the overall process, the calculation efficiency is improved by the fuzzy control method, and the calculation accuracy is improved by the LSTM neural network, so the technical solution of the present invention can take into account both efficiency and accuracy when realizing water supply control.

[0034] like Figure 2 As shown, an embodiment of the present invention provides an intelligent water supply control method. Compared with the above-mentioned embodiment, the intelligent water supply control method of this embodiment, before step S120, further includes:

[0035] Step S210: In the fuzzy control model, set the minimum value P of the water supply pressure of the urban water supply system. min and the maximum value P max , to determine the fuzzy domain of water supply pressure in urban water supply system.

[0036] Step S220, calculating the minimum water pressure P of the urban water supply system min and the maximum value P max The mean value P mid .

[0037] Step S230, setting a membership function for the water pressure of the urban water supply system Wherein, P is a variable representing the water supply pressure of the urban water supply system.

[0038] According to the technical solution of this embodiment, a fuzzy domain based on the mean and a simple Gaussian membership function are used, which not only ensures the expressiveness of the model, but also avoids overly complex model design and reduces the demand for computing resources. Using Gaussian distribution as a membership function can better simulate the uncertainty of the temperature parameter, and give a higher membership when the temperature is close to the middle value. As the temperature moves away from the middle value, the membership gradually decreases. Compared with the linear membership function, the membership function in the form of Gaussian distribution provides a nonlinear change trend, which can better capture the complexity of the temperature parameter changing over time.

[0039] like Figure 3 As shown, an embodiment of the present invention provides an intelligent water supply control method. Compared with the above-mentioned embodiment, the intelligent water supply control method of this embodiment, before step S120, further includes:

[0040] Step S310, detecting the geographical location of the city water supply system to determine the region where the city water supply system is located.

[0041] Step S320, query the historical maximum humidity H of the area where the urban water supply system is located hmax and the lowest humidity in history hmin .

[0042] Step S330, taking the minimum humidity of the city water supply system at the current time Take the maximum humidity of the city water supply system at the current time To determine the fuzzy domain of humidity in the city water supply system at the current time, where H bmin is the preset minimum basic humidity value, H bmax is the preset maximum basic humidity value, and ΔH is the preset humidity adjustment range.

[0043] According to the technical solution of this embodiment, by detecting the geographical location of the urban water supply system and determining the humidity fuzzy domain at the current time based on the historical highest humidity and the historical lowest humidity at the geographical location, the change of humidity at the pouring site can be more accurately reflected, thereby improving the adjustment accuracy of the water supply strategy. The setting of the fuzzy domain can handle the uncertainty of humidity, making the system more robust, and making reasonable water supply adjustments even in the face of humidity fluctuations.

[0044] Step S340, detecting the weather conditions in the area where the urban water supply system is located at the current time.

[0045] Step S350, setting the weather impact factor γ according to the weather conditions in the area where the urban water supply system is located at the current time wea .

[0046] Step S360, adjusting the minimum and maximum values ​​of humidity in the city water supply system at the current time according to the weather influencing factors. The maximum humidity of the city water supply system at the current time

[0047] According to the technical solution of this embodiment, by detecting the weather conditions in the area where the urban water supply system is located at the current time and setting the weather impact factor accordingly, the impact of weather changes on the humidity of the pouring site can be more accurately reflected, thereby improving the adjustment accuracy of the water supply strategy. The introduction of the weather impact factor can handle the impact of weather changes on humidity, making the system more robust and able to make reasonable water supply adjustments even in the face of weather changes.

[0048] An embodiment of the present invention provides an intelligent water supply control method. Compared with the above-mentioned embodiment, the intelligent water supply control method of this embodiment, before step S120, further includes:

[0049] (1) Multiple fuzzy rules in the fuzzy control model are recorded in a preset file, and the file is used to query the required fuzzy rules when using the fuzzy control model for calculation.

[0050] (2) Sort the multiple fuzzy calculation rules according to the pre-recorded usage counts of the multiple fuzzy rules.

[0051] (3) According to the order of the multiple fuzzy calculation rules, the positions of the multiple fuzzy rules in the file are adjusted.

[0052] According to the technical solution of this embodiment, by sorting the fuzzy rules and adjusting the position in the file, those frequently used fuzzy rules can be accessed first, thereby speeding up the response speed of the fuzzy control model. By sorting the fuzzy rules, the storage structure in the file can be optimized, unnecessary search and reading operations can be reduced, and storage space can be saved.

[0053] like Figure 4 As shown, an embodiment of the present invention provides a smart water supply control method. Compared with the above-mentioned embodiment, the smart water supply control method of this embodiment, before step S130, further includes:

[0054] Step S410, when training the LSTM neural network, set the unbalanced loss L1=w for the LSTM neural network e (y1-y2) 2 , where y1 is the water supply strategy output during LSTM neural network training, y2 is the target water supply strategy that should be output during LSTM neural network training, and w e It represents the impact level of the gap between the water supply strategy output by the LSTM neural network and the target water supply strategy that should be output on the occurrence rate of water supply accidents.

[0055] In this embodiment, by setting the unbalanced loss, the gap between the water supply strategy output by the LSTM neural network and the target water supply strategy can be more accurately measured, thereby improving the prediction accuracy of the model.

[0056] Step S420, setting regularization loss for LSTM neural network Among them, τ is the preset regularization coefficient, z is the total number of weights in the LSTM neural network, and w i is the i-th weight.

[0057] In this embodiment, by introducing regularization loss, the complexity of the model can be limited, overfitting can be prevented, and the generalization ability of the model can be improved.

[0058] Step S430, setting the loss function L=L1+L2 for the LSTM neural network.

[0059] Step S440, calculating the value of the loss function L(n) after the nth training of the LSTM neural network.

[0060] According to the technical solution of this embodiment, by introducing non-equilibrium loss and regularization loss and defining a comprehensive loss function, the degree of automation, accuracy and safety in the water supply process can be significantly improved, while reducing costs and energy consumption, which is of great significance to improving the stability and efficiency of urban water supply.

[0061] like Figure 5 As shown, an embodiment of the present invention provides a smart water supply control method. Compared with the above-mentioned embodiment, the smart water supply control method of this embodiment, before step S130, further includes:

[0062] Step S510, when the LSTM neural network is trained for the nth time, the current statistical period is calculated And round its value, where t is a positive integer.

[0063] Step S520: compare the current statistical period r with the recorded historical statistical period r h size.

[0064] Step S530: when the current statistical period r is greater than the historical statistical period r h When comparing the loss function L(n) after the nth training of the LSTM neural network with the pre-recorded historical loss function L h The size of the value.

[0065] In this embodiment, by regularly checking the improvement of the loss function and reducing the learning rate when the loss function value improves, the model training can be performed more efficiently, unnecessary training iterations can be reduced, and the training speed can be improved.

[0066] Step S540, the value of the loss function L(n) after the nth training of the LSTM neural network is less than the historical loss function L h When the value of is smaller, the learning rate ρ is reduced.

[0067] In this embodiment, by reducing the learning rate when the loss function value improves, the present invention can avoid oscillation when the model training enters the vicinity of the local minimum, which helps to prevent the occurrence of overfitting.

[0068] Step S550: Use the current statistical period r to update the historical statistical period r h , use the value of the loss function L(n) after the nth training of the LSTM neural network to update the historical loss function L h .

[0069] Wherein step S540 comprises:

[0070] (1) Calculate the target learning rate ρ tar =ρ0D 1+r , where D is between 0 and 1.

[0071] (2) Determine the target learning rate ρ tar Is the value lower than the preset minimum learning rate?

[0072] (3) When the judgment result is no, according to the target learning rate ρ tar Update the value of the learning rate ρ.

[0073] (4) When the judgment result is yes, the value of the learning rate ρ is updated according to the minimum learning rate.

[0074] According to the technical solution of this embodiment, by calculating the target learning rate and adjusting the learning rate in a timely manner, the model training can be performed more efficiently, unnecessary training iterations can be reduced, and the training speed can be improved. By setting the minimum learning rate limit, slow convergence caused by too small a learning rate during model training can be avoided, which helps prevent the occurrence of overfitting.

[0075] like Figure 6 As shown, in one embodiment of the present invention, a smart water supply control system is provided, comprising:

[0076] The environmental parameter collection module 610 collects the environmental parameters of the city water supply system at the current time, including the temperature, humidity, and water pressure of the city water supply system.

[0077] The water supply calculation module 620 calculates the environmental parameters of the urban water supply system at the current time according to a preset fuzzy control model to obtain the water supply regulation amount at the current time.

[0078] In this embodiment, the fuzzy control model can handle the uncertainty and nonlinear problems in the water supply regulation amount estimation process, and the fuzzy control calculation process is relatively fast.

[0079] The water supply strategy calculation module 630 inputs the environmental parameters of the city water supply system at the current time and the water supply regulation amount into a preset LSTM neural network. The LSTM neural network has a multi-layer neuron structure and is trained based on the environmental parameters of the city water supply system at historical time, the water supply regulation amount and the target water supply strategy through the water pump. For any weight w of the associated two adjacent layers of neurons in the LSTM neural network: and Wherein w(n), w(n+1), and w(n-1) are the values ​​of the weight w after the nth, n+1th, and n-1th trainings, respectively; ρ is a preset learning rate; E(n) is the value of the error function corresponding to the LSTM neural network after the nth training; θ is the partial differential symbol; α is the preset momentum term coefficient; L(n) is the value of the loss function of the LSTM neural network after the nth training; and β is a preset smoothing coefficient.

[0080] In this embodiment, the LSTM neural network can predict future water supply demand based on current environmental parameters by learning historical data, thereby adjusting the water supply strategy more accurately to ensure that the water supply volume matches the water supply pressure. The training of the LSTM neural network adopts an efficient update formula, and by introducing momentum terms and adaptive learning rate adjustment strategies, the convergence speed and computational efficiency of the model are improved, making the calculation of the water supply strategy faster.

[0081] The LSTM neural network is used to output the water supply strategy of the urban water supply system at the current time t, including the flow rate and lift of the pump.

[0082] The water supply module 640 controls the valve opening of the water pump according to the water supply strategy of the water pump so that the water pump supplies water to supply water in the city water supply system.

[0083] According to the technical solution of this embodiment, by adopting a preset fuzzy control model and LSTM neural network, automatic water supply control is realized, the need for manual intervention is reduced, and the workload of operators is reduced. By real-time monitoring and analysis of the environmental parameters of the urban water supply system, the water supply strategy can be adjusted in time to ensure that the water supply stability is in an ideal state during the water supply process, thereby improving the stability of urban water supply. In the overall process, the calculation efficiency is improved by the fuzzy control method, and the calculation accuracy is improved by the LSTM neural network, so the technical solution of the present invention can take into account both efficiency and accuracy when realizing water supply control.

[0084] The basic principles of the present application are described above in conjunction with specific embodiments. However, it should be noted that the advantages, strengths, effects, etc. mentioned in the present application are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. are required by each embodiment of the present application. In addition, the specific details disclosed above are only for the purpose of illustration and ease of understanding, not for limitation, and the above details do not limit the present application to being implemented by adopting the above specific details.

[0085] The block diagrams of the devices, apparatuses, equipment, and systems involved in this application are only illustrative examples and are not intended to require or imply that they must be connected, arranged, and configured in the manner shown in the block diagram. As will be appreciated by those skilled in the art, these devices, apparatuses, equipment, and systems can be connected, arranged, and configured in any manner. Words such as "including", "comprising", "having", etc. are open words, referring to "including but not limited to", and can be used interchangeably with them. The words "or" and "and" used here refer to the words "and / or" and can be used interchangeably with them, unless the context clearly indicates otherwise. The words "such as" used here refer to the phrase "such as but not limited to", and can be used interchangeably with them.

[0086] It should also be noted that in the apparatus, device and method of the present application, each component or each step can be decomposed and / or recombined. Such decomposition and / or recombination should be regarded as equivalent solutions of the present application.

[0087] The above description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other aspects without departing from the scope of the present application. Therefore, the present application is not intended to be limited to the aspects shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

[0088] The above description has been given for the purpose of illustration and description. In addition, this description is not intended to limit the embodiments of the present application to the forms disclosed herein. Although multiple example aspects and embodiments have been discussed above, those skilled in the art will recognize certain variations, modifications, changes, additions and sub-combinations thereof.

Claims

1. Intelligent water supply control method, characterized in that: include: Collecting environmental parameters of the city water supply system at the current time, including temperature, humidity, and water pressure of the city water supply system; According to a preset fuzzy control model, the environmental parameters of the urban water supply system at the current time are calculated to obtain the water supply regulation amount at the current time; The environmental parameters and water supply regulation of the city water supply system at the current time are input into a preset LSTM neural network, which has a multi-layer neuron structure and is trained based on the environmental parameters, water supply regulation of the city water supply system at historical time and the target water supply strategy through the water pump, wherein, for any weight w of the associated two adjacent layers of neurons in the LSTM neural network: and Where w(n), w(n+1), and w(n-1) are the values ​​of weight w after the nth, n+1th, and n-1th trainings, respectively; ρ is the preset learning rate; and E(n) is the value of the error function corresponding to the LSTM neural network after the nth training. is the partial differential symbol, α is the preset momentum term coefficient, L(n) is the value of the loss function after the nth training of the LSTM neural network, and β is the preset smoothing coefficient; Outputting the water supply strategy of the water pump of the city water supply system at the current time t through the LSTM neural network, including the flow rate and lift of the water pump; According to the water supply strategy of the water pump, the valve opening of the water pump is controlled to make the water pump supply water to supply water in the city water supply system.

2. The intelligent water supply control method according to claim 1, characterized in that: Before calculating the environmental parameters of the urban water supply system at the current time according to the preset fuzzy control model, the method further includes: In the fuzzy control model, the minimum value P of the water supply pressure of the urban water supply system is set min and the maximum value P max , to determine the fuzzy domain of the water supply pressure of the urban water supply system; Calculate the minimum water pressure P of the city water supply system min and the maximum value P max The mean value P mid ; Setting a membership function for the water supply pressure of the urban water supply system Wherein, P is a variable representing the water supply pressure of the urban water supply system.

3. The intelligent water supply control method according to claim 1, characterized in that: Before calculating the environmental parameters of the urban water supply system at the current time according to the preset fuzzy control model, the method further includes: Detecting the geographical location of the city water supply system to determine the area where the city water supply system is located; Query the historical maximum humidity H of the area where the city water supply system is located hmax and the lowest humidity in history hmin ; Take the minimum value of humidity of the city water supply system at the current time Take the maximum value of the humidity of the city water supply system at the current time To determine the fuzzy domain of humidity of the city water system at the current time, where H bmin is the preset minimum basic humidity value, H bmax is the preset maximum basic humidity value, and ΔH is the preset humidity adjustment range.

4. The intelligent water supply control method according to claim 1, characterized in that: Also includes: Detecting the weather conditions in the area where the city water supply system is located at the current time; The weather influence factor γ is set according to the weather conditions in the area where the urban water supply system is located at the current time. wea ; The minimum and maximum values ​​of the humidity of the city water supply system at the current time are adjusted according to the weather influencing factor. The minimum value of the humidity of the city water supply system at the current time is The maximum humidity of the city water supply system at the current time 5. The intelligent water supply control method according to claim 1, characterized in that: Before calculating the environmental parameters of the urban water supply system at the current time according to the preset fuzzy control model, the method further includes: Recording a plurality of fuzzy rules in the fuzzy control model in a preset file, wherein the file is used to query required fuzzy rules when using the fuzzy control model for calculation; sorting the plurality of fuzzy calculation rules according to the pre-recorded usage counts of the plurality of fuzzy rules; According to the order of the multiple fuzzy calculation rules, the positions of the multiple fuzzy rules in the file are adjusted.

6. The intelligent water supply control method according to claim 1, characterized in that: Before inputting the environmental parameters and water supply regulation amount of the city water supply system at the current time into the preset LSTM neural network, it also includes: When training the LSTM neural network, an unbalanced loss L1=w is set for the LSTM neural network. e (y1-y2) 2 , where y1 is the water supply strategy output during the LSTM neural network training, y2 is the target water supply strategy that should be output during the LSTM neural network training, and w e Indicates the impact level of the gap between the water supply strategy output by the LSTM neural network and the target water supply strategy that should be output on the occurrence rate of water supply accidents; Set the regularization loss for the LSTM neural network Wherein, τ is the preset regularization coefficient, z is the total number of weights in the LSTM neural network, and w i is the i-th weight; Setting a loss function L=L1+L2 for the LSTM neural network; Calculate the value of the loss function L(n) after the nth training of the LSTM neural network.

7. The intelligent water supply control method according to claim 6, characterized in that: Before inputting the environmental parameters and water supply regulation amount of the city water supply system at the current time into the preset LSTM neural network, it also includes: When the LSTM neural network is trained for the nth time, the current statistical period is calculated And round its value, where t is a positive integer; Compare the current statistical period r with the recorded historical statistical period r h size; When the current statistical period r is greater than the historical statistical period r h When the loss function L(n) of the LSTM neural network after the nth training is compared with the pre-recorded historical loss function L h The size of the value; The value of the loss function L(n) after the nth training of the LSTM neural network is less than the historical loss function L h When the value is , the learning rate ρ is reduced; Use the current statistical period r to update the historical statistical period r h , use the value of the loss function L(n) after the nth training of the LSTM neural network to update the historical loss function L h .

8. The intelligent water supply control method according to claim 7, characterized in that: The step of reducing the learning rate ρ comprises: Calculate the target learning rate ρ tar =ρ0D 1+r , where D takes values ​​between 0 and 1; Determine the target learning rate ρ tar Is the value of lower than the preset minimum learning rate? When the judgment result is no, according to the target learning rate ρ tar Update the value of the learning rate ρ with the value of When the judgment result is yes, the value of the learning rate ρ is updated according to the minimum learning rate.

9. Intelligent water supply control system, characterized in that: include: An environmental parameter collection module collects environmental parameters of the city water supply system at the current time, including temperature, humidity, and water pressure of the city water supply system; A water supply amount calculation module calculates the environmental parameters of the city water supply system at the current time according to a preset fuzzy control model to obtain the water supply adjustment amount at the current time; The water supply strategy calculation module inputs the environmental parameters of the city water supply system at the current time and the water supply regulation amount into a preset LSTM neural network. The LSTM neural network has a multi-layer neuron structure and is trained based on the environmental parameters of the city water supply system at historical time, the water supply regulation amount and the target water supply strategy through the water pump. For any weight w of the associated two adjacent layers of neurons in the LSTM neural network: and Where w(n), w(n+1), and w(n-1) are the values ​​of weight w after the nth, n+1th, and n-1th trainings, respectively; ρ is the preset learning rate; and E(n) is the value of the error function corresponding to the LSTM neural network after the nth training. is the partial differential symbol, α is the preset momentum term coefficient, L(n) is the value of the loss function after the nth training of the LSTM neural network, and β is the preset smoothing coefficient; the water supply strategy of the water pump of the urban water supply system at the current time t is output through the LSTM neural network, including the flow rate and head of the water pump; The water supply module controls the valve opening of the water pump according to the water supply strategy of the water pump so that the water pump supplies water to supply water in the city water supply system.

Images