CN111624874A - Pump station cluster intelligent prediction method and system for urban sewage treatment - Google Patents

Pump station cluster intelligent prediction method and system for urban sewage treatment Download PDF

Info

Publication number
CN111624874A
CN111624874A CN202010516998.6A CN202010516998A CN111624874A CN 111624874 A CN111624874 A CN 111624874A CN 202010516998 A CN202010516998 A CN 202010516998A CN 111624874 A CN111624874 A CN 111624874A
Authority
CN
China
Prior art keywords
data
pump station
pump
liquid level
stations
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
Application number
CN202010516998.6A
Other languages
Chinese (zh)
Other versions
CN111624874B (en
Inventor
方超
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Zhejiang Chaofan Environmental Technology Co ltd
Original Assignee
Zhejiang Chaofan Environmental Technology Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Zhejiang Chaofan Environmental Technology Co ltd filed Critical Zhejiang Chaofan Environmental Technology Co ltd
Priority to CN202010516998.6A priority Critical patent/CN111624874B/en
Publication of CN111624874A publication Critical patent/CN111624874A/en
Application granted granted Critical
Publication of CN111624874B publication Critical patent/CN111624874B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
    • G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
    • G05B13/0265—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion
    • G05B13/027—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric the criterion being a learning criterion using neural networks only
    • 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
    • F04B49/065—Control using electricity and making use of computers

Landscapes

  • Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Computer Hardware Design (AREA)
  • Mechanical Engineering (AREA)
  • General Engineering & Computer Science (AREA)
  • Feedback Control In General (AREA)

Abstract

The invention relates to the field of remote monitoring, and provides a pump station cluster intelligent prediction method for urban sewage treatment, which comprises the steps of acquiring spatial data, time data and additional influence data of a plurality of adjacent pump stations, inputting the spatial data into a multilayer convolutional neural network, inputting the time data into a cyclic neural network model, inputting the additional influence data into two layers of full-connection layers respectively to obtain corresponding output results, carrying out weighted calculation on the data to predict pump station liquid level data, carrying out comparison calculation feedback on actual data for observing the pump station liquid level, and adjusting weight, so that the method can more efficiently and accurately predict the output of the pump station liquid level result data and further trigger the working efficiency of a driving motor of the corresponding pump station, and the invention also provides a system which can predict the change rule of the pump station liquid level in a period of the future, the urban sewage treatment pump station of the intelligent cooperative dispatching cluster is high in efficiency and intelligent degree.

Description

Pump station cluster intelligent prediction method and system for urban sewage treatment
Technical Field
The invention relates to the field of remote monitoring, in particular to a pump station cluster intelligent prediction method and a pump station cluster intelligent prediction system for urban sewage treatment.
Background
A large number of pump stations for sewage lifting are distributed in a city, and the existence of the pump stations can convey domestic sewage, industrial sewage and the like in the city to a sewage treatment plant for purification treatment. The sewage pump station is an important component of a sewage system, and has the characteristics of continuous water flow, small water flow and the like, and is widely applied to the urban sewage treatment system.
CN205942426U discloses a sewage scheduling system for a sewage pipe network pump station, which comprises a regulating reservoir pump station unit and a control system; the control system comprises a pump station field remote centralized control device, a wireless data acquisition and processing device and a remote monitoring or dispatching center. The equalizing basin pump station unit comprises a primary equalizing basin and a secondary equalizing basin, wherein the output end of the primary equalizing basin is connected to the input end of the secondary equalizing basin, and the output end of the secondary equalizing basin is connected to the input end of the sewage treatment plant.
At present, sewage lifting pump stations in cities still operate in an independent mode, the pump stations are mainly distributed in the cities in an independent mode, each pump station works in the independent mode, the system optimization of the pump stations only stops in the aspect of networking monitoring of a single pump station, because the pump stations are connected in series or in parallel, the traditional monitoring system does not have the scheduling capability of cooperatively processing a plurality of pump stations, the sewage conveying among the pump stations is completed in a relay mode, if the problem of extreme weather or uncoordinated work of adjacent pump stations is caused, the liquid level in a relay pump station is easy to rise suddenly, and finally, part of the pump stations are overloaded to work, and the internal sewage cannot be extracted in time to present a blowout state.
Disclosure of Invention
Long-term practice shows that because sewage lifting pump stations in cities operate independently, the traditional method and the traditional system only optimize and monitor a single pump station, however, the pump stations are connected in series or in parallel, sewage conveying is completed between the pump stations in a relay mode, the traditional monitoring system does not have the scheduling capability of cooperatively processing a plurality of pump stations, the forecasting capability is lacked, and if the problem of extreme weather or uncoordinated work of adjacent pump stations, the liquid level in the relay pump station is easy to rise suddenly, and finally, part of the pump stations are overloaded to work, and cannot extract internal sewage in time to present a 'blowout' state.
In view of the above, the present invention aims to provide an intelligent prediction method for a pump station cluster for urban sewage treatment, so as to solve the problems that the current urban sewage treatment pump station cluster has insufficient cooperative processing capability among a plurality of pump stations and lacks intelligent prediction, and the intelligent prediction method for the pump station cluster for urban sewage treatment comprises:
step S1, acquiring spatial data, time data and additional influence data of a plurality of adjacent pump stations, wherein the spatial data comprises liquid level data of the adjacent pump stations, flow data of the adjacent pump stations and capacity of the pump stations; the time data comprises historical liquid level data and historical flow data; the additional impact data comprises weather status data;
step S2, inputting the acquired space data into multilayer convolution neural network, outputting the result data
Figure BDA0002530469120000021
Figure BDA0002530469120000022
Wherein the content of the first and second substances,
Figure BDA0002530469120000023
inputting tensors for the convolution layers, wherein the tensors comprise tensors formed by liquid level data of adjacent pump stations, flow data of adjacent pump stations and capacity of the pump stations;
Figure BDA0002530469120000024
in order to be the parameters of the convolution kernel,
Figure BDA0002530469120000025
for the bias parameter, f is the activation function,
Figure BDA0002530469120000026
outputting a tensor for the convolutional layer, a convolution operation; the convolutional neural network comprises a deep residual network ResNet;
step S3, inputting the acquired time data into a recurrent neural network model, and outputting t time point prediction pump station liquid level data
Figure BDA0002530469120000027
Wherein the historical liquid level data of the pump station at the past t-1 time point is
X＝{X1，X2，X3，...，Xt-1}；
State h of hidden layer output in t time point cyclic neural network modeltIs composed of
ht＝σh(wxhXt+whhht-1+bh)；
Wherein σhTo activate a function, wxhFor the weight parameter between the input layer and the hidden layer, whhIs a weight parameter between the hidden layer and the hidden layer, ht-1At the state of the previous moment, bhIs an offset; the basic unit in the recurrent neural network model is an LSTM network unit;
step S4, setting the size of the characteristic tensor which can be extracted by the additional influence data input into two fully-connected layers which are the same as the input sample tensor, and outputting the tensor
Figure BDA0002530469120000031
Step S5, calculating the output data of step S2, step S3 and step S4 to obtain result data of liquid level of the prediction pump station
Figure BDA0002530469120000032
Figure BDA0002530469120000033
Wherein, we、wr、wlFor the weight matrix that can be obtained in the model training,. is a matrix multiplication; then will be
Figure BDA0002530469120000034
Inputting the result into an activation function tanh for calculation to obtain result data of liquid level of a prediction pump station
Figure BDA0002530469120000035
Figure BDA0002530469120000036
Step S6, will
Figure BDA0002530469120000037
And the actual observation result XtComparing and calculating a prediction error E;
Figure BDA0002530469120000038
wherein N is tensor XtNumber of elements of (2), xiIs the observed value of the liquid level of the pump station i,
Figure BDA0002530469120000039
the predicted value of the liquid level of the pump station i is obtained;
step S7, when the prediction error is less than the predetermined value
Figure BDA00025304691200000310
And generating a trigger signal by the capacity data of the pump station, and sending the trigger signal to the corresponding pump station PLC control unit for adjusting the running power of the water pump.
Preferably, in step S2, the multi-layer convolutional neural network is a deep learning model.
Preferably, in step S3, the recurrent neural network model outputs the result ytIs composed of
yt＝σy(whoht+bo)；
Wherein σyFor the output layer activation function, whoTo hide the weighting parameters of the output layer from the layer, boIndicating the bias of the output layer.
Preferably, in step S3, the input parameters of the recurrent neural network model with LSTM network unit as the basic unit include historical liquid level data at t-1 time before the pump station, and the pump station liquid level data at t time is output.
Preferably, in step S6, w can be feedback-adjusted according to the magnitude of the prediction error Ee、wr、wlStopping adjusting w when the prediction error E value is less than the preset valuee、wr、wlAnd (5) completing the training of the model.
In order to better execute the method, the invention also provides a system for executing the intelligent prediction method of the pump station cluster for urban sewage treatment, wherein the system comprises a computing processing unit, a plurality of pump stations, a sensor assembly and a PLC control unit of the intelligent prediction method of the pump station cluster for urban sewage treatment;
the sensor assembly is used for acquiring running state data and sewage data of a plurality of pump stations, and the running state data and the sewage data comprise adjacent pump station liquid level data, adjacent pump station flow data, historical liquid level data, historical flow data and additional influence data;
the calculation processing unit is connected with the sensor assembly through a network, receives and processes multi-source information data generated by the sensor assembly, calculates and processes the data by adopting a pump station cluster intelligent prediction method for urban sewage treatment, and generates a trigger signal for driving and adjusting the running power of a plurality of pump stations to the PLC control unit;
and the PLC control unit is used for receiving the trigger signal generated by the calculation processing unit and generating a specific execution signal for adjusting the running power of a plurality of pump stations to the drive motor of the pump station.
Preferably, the system further comprises a cloud management platform, wherein the cloud management platform is used for storing a plurality of pump station running state data and sewage data collected by the sensor assembly, storing and displaying pump station distribution position data and pump station state data, recording the running state of a pump station cluster, and executing pump station video retrieval, remote pump station startup and shutdown and pump station parameter change operations;
the cloud management platform comprises the computing processing unit, and the computing processing unit is used for computing and processing the time data, the space data and the additional influence data acquired by the sensor assembly.
Preferably, the system further comprises a video monitoring system used for acquiring video data of the surrounding environment of the pump station, and the video monitoring system is connected with the cloud management platform through a video network.
Preferably, the pump station includes sewage lifting pump station, sewage lifting pump station include switch board, sewage storage jar, water pump and sensor assembly, the switch board with the water pump the sensor assembly electricity is connected, sewage storage jar with the water pump is connected, sensor assembly with sewage storage jar fixed connection.
According to another aspect of the embodiments of the present invention, there is provided a storage medium, the storage medium including a stored program, wherein when the program runs, a device in which the storage medium is located is controlled to execute the above method.
According to the pump station cluster intelligent prediction method for urban sewage treatment, the spatial data, the time data and the additional influence data of a plurality of adjacent pump stations are obtained, the spatial data are input into a multilayer convolutional neural network, the time data are input into a cyclic neural network model, the additional influence data are respectively input into two full-connection layers to obtain corresponding output results, three groups of data are weighted to calculate and predict pump station liquid level result data, actual data for observing the liquid level of the pump station are compared, calculated and fed back, the weight is further adjusted, so that the method can more efficiently and accurately predict the output of the pump station liquid level result data and further trigger the working efficiency of a driving motor of the corresponding pump station, in order to better execute the method, the embodiment of the invention also provides a system, and the method and the system can predict the liquid level change rule of the pump station in a period of the future, the urban sewage treatment pump station of intelligent cooperative scheduling cluster, it is efficient, intelligent degree is high, be unlikely to appear that some pump station work is overloaded, can't in time extract inside sewage and the problem that appears "blowout" state that leads to.
Additional features and advantages of the invention will be set forth in the detailed description which follows.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate an embodiment of the invention and, together with the description, serve to explain the invention and not to limit the invention. In the drawings:
FIG. 1 is a flow chart of an intelligent prediction method for a pump station cluster for urban sewage treatment according to an embodiment of the present invention;
FIG. 2 is a process diagram of a convolutional neural network in accordance with an embodiment of the present invention;
FIG. 3 is a ResNet network element structure according to an embodiment of the present invention;
FIG. 4 is a recurrent neural network model in accordance with an embodiment of the present invention;
FIG. 5 is a block diagram of an LSTM cell in accordance with an embodiment of the present invention;
FIG. 6 is a prediction model for spatio-temporal data fusion according to an embodiment of the present invention;
FIG. 7 is a block diagram of a system for performing a method in accordance with one embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the present invention, are given by way of illustration and explanation only, not limitation.
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It should be noted that the terms "first," "second," "third," and the like in the description and in the claims, and in the drawings, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged under appropriate circumstances in order to facilitate the description of the embodiments of the invention herein. Furthermore, the terms "comprises," "comprising," and "having," and any variations thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
The system aims to solve the problems that in the prior art, only a single pump station is optimized and monitored in the process of treating urban sewage, however, the pump stations are connected in series or in parallel, sewage conveying is completed in the form of relay among the pump stations, the system does not have the scheduling capability of cooperatively treating a plurality of pump stations, the system lacks of prediction capability, and if the problem of extreme weather or uncoordinated work of adjacent pump stations, the liquid level in the relay pump station is easy to rise suddenly, and finally, part of the pump stations are overloaded to work, and the internal sewage cannot be extracted in time to present a 'blowout' state. As shown in fig. 1 to 6, the present invention provides a pump station cluster intelligent prediction method for municipal sewage treatment, and as shown in fig. 1, a flow chart of a pump station cluster intelligent prediction method for municipal sewage treatment according to an embodiment of the present invention, where the pump station cluster intelligent prediction method for municipal sewage treatment includes:
step S1, acquiring spatial data, time data and additional influence data of a plurality of adjacent pump stations, wherein the spatial data comprises liquid level data of the adjacent pump stations, flow data of the adjacent pump stations and capacity of the pump stations; the time data comprises historical liquid level data and historical flow data; the additional impact data comprises weather status data;
step S2, inputting the acquired space data into multilayer convolution neural network, outputting the result data
Figure BDA0002530469120000071
Figure BDA0002530469120000072
Wherein the content of the first and second substances,
Figure BDA0002530469120000073
inputting tensors for the convolution layers, wherein the tensors comprise tensors formed by liquid level data of adjacent pump stations, flow data of adjacent pump stations and capacity of the pump stations;
Figure BDA0002530469120000074
in order to be the parameters of the convolution kernel,
Figure BDA0002530469120000075
for the bias parameter, f is the activation function,
Figure BDA0002530469120000076
outputting a tensor for the convolutional layer, a convolution operation; the convolutional neural network comprises a deep residual network ResNet;
step S3, inputting the acquired time data into a recurrent neural network model, and outputting t time point prediction pump station liquid level data
Figure BDA0002530469120000077
Wherein the historical liquid level data of the pump station at the past t-1 time point is
X＝{X1，X2，X3，...，Xt-1}；
State h of hidden layer output in t time point cyclic neural network modeltIs composed of
ht＝σh(wxhXt+whhht-1+bh)；
Wherein σhTo activate a function, wxhFor the weight parameter between the input layer and the hidden layer, whhIs a weight parameter between the hidden layer and the hidden layer, ht-1At the state of the previous moment, bhIs an offset; the basic unit in the recurrent neural network model is an LSTM network unit;
step S4, setting the size of the characteristic tensor which can be extracted by the additional influence data input into two fully-connected layers which are the same as the input sample tensor, and outputting the tensor
Figure BDA0002530469120000081
Step S5, calculating the output data of step S2, step S3 and step S4 to obtain result data of liquid level of the prediction pump station
Figure BDA0002530469120000082
Figure BDA0002530469120000083
Wherein, we、wr、wlFor the weight matrix that can be obtained in the model training,. is a matrix multiplication; then will be
Figure BDA0002530469120000084
Inputting the result into an activation function tanh for calculation to obtain result data of liquid level of a prediction pump station
Figure BDA0002530469120000085
Figure BDA0002530469120000086
Step S6, will
Figure BDA0002530469120000087
And the actual observation result XtComparing and calculating a prediction error E;
Figure BDA0002530469120000088
wherein N is tensor XtNumber of elements of (2), xiIs the observed value of the liquid level of the pump station i,
Figure BDA0002530469120000089
the predicted value of the liquid level of the pump station i is obtained;
step S7, when the prediction error is less than the predetermined value
Figure BDA00025304691200000810
And capacity data generation touch of pump stationAnd sending a signal to the corresponding pump station PLC control unit for adjusting the running power of the water pump.
The pump station cluster intelligent prediction method for urban sewage treatment provided by the embodiment of the invention comprises the steps of acquiring spatial data, time data and extra influence data of a plurality of adjacent pump stations, inputting the spatial data into a multilayer convolutional neural network, inputting the time data into a cyclic neural network model, inputting the extra influence data into two full-connection layers respectively to obtain corresponding output results, carrying out weighting calculation on three groups of data to predict pump station liquid level result data, carrying out comparison calculation feedback on actual data for observing the liquid level of the pump station, further adjusting the weight, thus leading the method to more efficiently and accurately predict the output of the pump station liquid level result data and further triggering the working efficiency of a driving motor of the corresponding pump station, being capable of predicting the liquid level change rule of the pump station in a period of time in the future and intelligently and cooperatively scheduling the clustered urban sewage treatment pump stations, high efficiency, intelligent degree is high, is unlikely to appear that some pump station work is transshipped, can't in time extract inside sewage and the problem that presents "blowout" state that leads to.
Because the urban range is very large, the high semantic features can be obtained only by convolution of dozens of layers or even hundreds of layers when the sewage related features of the urban pumping station are extracted, however, the network can be degraded along with the increase of the number of layers of the convolutional neural network.
In order to convert originally low-semantic spatial information into high-semantic features by using a Convolutional Neural Network (CNN), as shown in fig. 2-3, in a more preferred case of the present invention, the multilayer Convolutional Neural network is a deep learning model in step S2, and in a more preferred case, as shown in fig. 3, a ResNet network unit structure.
In the preferred case of the invention, the historical liquid level data of the current pump station in the past period can be used for predicting the liquid level of the pump station in the time dimension, and a recurrent neural network (Recurr) which is good at processing the time dimension data is adoptedent neural network, RNN), the output of which is determined by the state of the previous time and the current feature input, the model structure is as shown in fig. 4, in order to better process the time series data, in a preferred case of the present invention, the recurrent neural network model outputs the result y in step S3tIs composed of
yt＝σy(whoht+bo)；
Wherein σyFor the output layer activation function, whoTo hide the weighting parameters of the output layer from the layer, boIndicating the bias of the output layer.
Because the traditional RNN model has limited Memory capacity, in order to find a change rule from Long-term liquid level historical data, the invention provides a Long short term Memory model (LSTM) capable of realizing Long term Memory. The basic unit is shown in FIG. 5, and mainly comprises an input gate itForgetting door ftAnd an output gate otThe input gate and the output gate are used for controlling input and output of information flow, the forgetting gate is used for controlling the state of the previous moment, and long-term memory is realized through the three gate operations.
The parameters in the LSTM structure are shown in the following formula,
Figure BDA0002530469120000101
wherein c istIndicates the cell state at time t, htOutput representing hidden layer, bf、bi、bo、bcTo be biased, wxf、whf、wcf、wxi、whi、wci、wxo、who、wco、wxc、whcAre weight coefficients.
Then, the LSTM base unit shown in fig. 5 replaces the base unit a of the RNN in fig. 4, so that the LSTM network is obtained. And inputting the historical liquid level data of the pump station at the previous t-1 moment into the LSTM model, so that the liquid level of the pump station at the t moment can be predicted.
In order to better calculate and obtain the current pump station liquid level data according to the historical liquid level data of the pump station at the previous t-1 moment, in the preferable condition of the invention, in the step S3, the input parameters of the recurrent neural network model of which the basic unit is an LSTM network unit comprise the historical liquid level data of the pump station at the previous t-1 moment, and the pump station liquid level data of the t moment is output.
In order to better adjust and train the model according to the actually observed pump station liquid level data so that the model can more accurately predict the corresponding liquid level change of the pump station, in the preferred case of the invention, in the step S6, the w can be adjusted in a feedback manner according to the prediction error E valuee、wr、wlStopping adjusting w when the prediction error E value is less than the preset valuee、wr、wlAnd (5) completing the training of the model.
In step S6, as shown in fig. 6, the output results calculated in step S2, step S3, and step S4 are subjected to calculation processing. For example, in a more preferred aspect of the invention,
in step S4 of the present invention, for additional influencing factor calculation, such as weather conditions, the output tensor of the data after passing through two fully connected layers
Figure BDA0002530469120000111
Where the main role of Conv2 is to set the size of the additional factor extracted feature tensor to be the same as the input sample tensor.
In step S2 of the present invention, the prediction of spatial data is performed by first inputting the related data of the neighboring pump stations into a convolutional layer Conv1, connecting S residual units under the convolutional layer, and finally obtaining the data of the model through Conv2
Figure BDA0002530469120000112
And setting the tensor size before and after the convolution operation to be unchanged in the model convolution process.
In the step S3 of the invention, time data prediction is carried out, historical liquid level data of the current pump station is taken and input into an LSTM model, and an output result is obtained
Figure BDA0002530469120000113
The output tensor is the same size as the input tensor. Under the more preferable condition of the invention, the model extracts the key data frame as the training data instead of all the historical data as the training data, so that the scale of the model can be greatly reduced, and the feasibility of large-scale urban calculation is increased.
The embodiment of the invention also provides a system for executing the intelligent prediction method of the pump station cluster for urban sewage treatment, which comprises a calculation processing unit, a plurality of pump stations, a sensor assembly and a PLC control unit, wherein the calculation processing unit is used for the intelligent prediction method of the pump station cluster for urban sewage treatment;
the sensor assembly is used for acquiring running state data and sewage data of a plurality of pump stations, and the running state data and the sewage data comprise adjacent pump station liquid level data, adjacent pump station flow data, historical liquid level data, historical flow data and additional influence data;
the calculation processing unit is connected with the sensor assembly through a network, receives and processes multi-source information data generated by the sensor assembly, calculates and processes the data by adopting a pump station cluster intelligent prediction method for urban sewage treatment, and generates a trigger signal for driving and adjusting the running power of a plurality of pump stations to the PLC control unit;
and the PLC control unit is used for receiving the trigger signal generated by the calculation processing unit and generating a specific execution signal for adjusting the running power of a plurality of pump stations to the drive motor of the pump station.
The system for executing the pump station cluster intelligent prediction method for urban sewage treatment provided by the embodiment of the invention obtains spatial data, time data and extra influence data of a plurality of adjacent pump stations, inputs the spatial data into a multilayer convolution neural network, inputs the time data into a circulation neural network model, respectively inputs the extra influence data into two layers of full-connection layers to obtain corresponding output results, performs weighting calculation on three groups of data to predict pump station liquid level result data, performs comparison calculation feedback on actual data for observing the pump station liquid level, further adjusts the weight, so that the method can more efficiently and accurately predict the output of the pump station liquid level result data and further trigger the working efficiency of a driving motor of the corresponding pump station, can predict the pump station liquid level change rule in a period of time in the future and intelligently and cooperatively schedule the clustered urban sewage treatment pump stations, high efficiency and high intelligent degree, and is unlikely to overload the work of part of pump stations.
In a preferred case of the present invention, as shown in fig. 7, in the preferred case of the present invention, the system further includes a cloud management platform, where the cloud management platform is configured to store a plurality of pump station operating state data and sewage data acquired by the sensor assembly, store and display pump station distribution position data and pump station state data, record an operating state of a pump station cluster, and execute pump station video retrieval, remote pump station startup and shutdown, and pump station parameter change operations;
in order to better process and calculate multi-source data acquired by the sensor assembly, under the preferable condition of the invention, the cloud management platform comprises the calculation processing unit, and the time data, the space data and the additional influence data acquired by the sensor assembly are calculated and processed, under the more preferable condition, the space data comprises adjacent pump station liquid level data, adjacent pump station flow data and pump station capacity; the time data comprises historical liquid level data and historical flow data; the additional impact data includes weather status data.
In order to better acquire the data of the surrounding environment of the pump station, the system preferably further comprises a video monitoring system for acquiring the video data of the surrounding environment of the pump station, and the video monitoring system is connected with the cloud management platform through a video network.
In order to enable the sensor assembly to acquire real-time data, under the optimal condition, the pump station comprises a sewage lifting pump station, the sewage lifting pump station comprises a control cabinet, a sewage storage tank, a water pump and the sensor assembly, the control cabinet is electrically connected with the water pump and the sensor assembly, the sewage storage tank is connected with the water pump, and the sensor assembly is fixedly connected with the sewage storage tank.
The embodiment of the invention also provides a storage medium, which comprises a stored program, wherein when the program runs, the device where the storage medium is located is controlled to execute the method.
It should be noted that, for simplicity of description, the above-mentioned method embodiments are described as a series of acts or combination of acts, but those skilled in the art will recognize that the present invention is not limited by the order of acts, as some steps may occur in other orders or concurrently in accordance with the invention. Further, those skilled in the art should also appreciate that the embodiments described in the specification are preferred embodiments and that the acts and modules referred to are not necessarily required by the invention.
In the foregoing embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus can be implemented in other manners. For example, the above-described embodiments of the apparatus are merely illustrative, and for example, the division of the units is only one type of division of logical functions, and there may be other divisions when actually implementing, for example, a plurality of units or components may be combined or may be integrated into another system, or some features may be omitted, or not implemented. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection of some interfaces, devices or units, and may be an electric or other form.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a mobile terminal, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a removable hard disk, a magnetic or optical disk, and other various media capable of storing program codes.
The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. The intelligent prediction method for the pump station cluster for the urban sewage treatment is characterized by comprising the following steps of:
step S1, acquiring spatial data, time data and additional influence data of a plurality of adjacent pump stations, wherein the spatial data comprises liquid level data of the adjacent pump stations, flow data of the adjacent pump stations and capacity of the pump stations; the time data comprises historical liquid level data and historical flow data; the additional impact data comprises weather status data;
step S2, inputting the acquired space data into multilayer convolution neural network, outputting the result data
Figure FDA0002530469110000011
Figure FDA0002530469110000012
Wherein the content of the first and second substances,
Figure FDA0002530469110000013
inputting tensors for the convolution layers, wherein the tensors comprise tensors formed by liquid level data of adjacent pump stations, flow data of adjacent pump stations and capacity of the pump stations;
Figure FDA0002530469110000014
in order to be the parameters of the convolution kernel,
Figure FDA0002530469110000015
for the bias parameter, f is the activation function,
Figure FDA0002530469110000016
outputting a tensor for the convolutional layer, a convolution operation; the convolutional neural network comprises a deep residual network ResNet;
step S3, inputting the acquired time data into a recurrent neural network model, and outputting t time point prediction pump station liquid level data
Figure FDA0002530469110000017
Wherein the historical liquid level data of the pump station at the past t-1 time point is
X＝{X1，X2，X3，...，Xt-1}；
State h of hidden layer output in t time point cyclic neural network modeltIs ht＝σh(wxhXt+whhht-1+bh)；
Wherein σhTo activate a function, wxhFor the weight parameter between the input layer and the hidden layer, whhTo hideWeight parameter between layer and hidden layer, ht-1At the state of the previous moment, bhIs an offset; the basic unit in the recurrent neural network model is an LSTM network unit;
step S4, setting the size of the characteristic tensor which can be extracted by the additional influence data input into two fully-connected layers which are the same as the input sample tensor, and outputting the tensor
Figure FDA0002530469110000018
Step S5, calculating the output data of step S2, step S3 and step S4 to obtain result data of liquid level of the prediction pump station
Figure FDA0002530469110000019
Figure FDA00025304691100000110
Wherein, we、wr、wlFor the weight matrix that can be obtained in the model training,. is a matrix multiplication; then will be
Figure FDA0002530469110000021
Inputting the result into an activation function tanh for calculation to obtain result data of liquid level of a prediction pump station
Figure FDA0002530469110000022
Figure FDA0002530469110000023
Step S6, will
Figure FDA0002530469110000024
And the actual observation result XtComparing and calculating a prediction error E;
Figure FDA0002530469110000025
wherein N is tensor XtNumber of elements of (2), xiIs the observed value of the liquid level of the pump station i,
Figure FDA0002530469110000026
the predicted value of the liquid level of the pump station i is obtained;
step S7, when the prediction error is less than the predetermined value
Figure FDA0002530469110000027
And generating a trigger signal by the capacity data of the pump station, and sending the trigger signal to the corresponding pump station PLC control unit for adjusting the running power of the water pump.
2. The pump station cluster intelligent prediction method for municipal sewage treatment according to claim 1, wherein in step S2, the multilayer convolutional neural network is a deep learning model.
3. The pump station cluster intelligent prediction method for municipal sewage treatment according to claim 1, wherein in step S3, the recurrent neural network model outputs result ytIs composed of
yt＝σy(whoht+bo)；
Wherein σyFor the output layer activation function, whoTo hide the weighting parameters of the output layer from the layer, boIndicating the bias of the output layer.
4. The pump station cluster intelligent prediction method for municipal sewage treatment according to claim 1, wherein in step S3, the input parameters of the recurrent neural network model with the basic unit being an LSTM network unit include historical liquid level data at t-1 time before the pump station, and the pump station liquid level data at t time is output.
5. For municipal sewage treatment according to any of claims 1 to 4The pump station cluster intelligent prediction method is characterized in that in step S6, w can be fed back and adjusted according to the prediction error E valuee、wr、wlStopping adjusting w when the prediction error E value is less than the preset valuee、wr、wlAnd (5) completing the training of the model.
6. A system, characterized in that the system comprises a calculation processing unit, a plurality of pump stations, a sensor assembly and a PLC control unit for executing the pump station cluster intelligent prediction method for urban sewage treatment according to any claim 1-5;
the sensor assembly is used for acquiring running state data and sewage data of a plurality of pump stations, and the running state data and the sewage data comprise adjacent pump station liquid level data, adjacent pump station flow data, historical liquid level data, historical flow data and additional influence data;
the calculation processing unit is connected with the sensor assembly through a network, receives and processes multi-source information data generated by the sensor assembly, calculates and processes the data by adopting a pump station cluster intelligent prediction method for urban sewage treatment, and generates a trigger signal for driving and adjusting the running power of a plurality of pump stations to the PLC control unit;
and the PLC control unit is used for receiving the trigger signal generated by the calculation processing unit and generating a specific execution signal for adjusting the running power of a plurality of pump stations to the drive motor of the pump station.
7. The system according to claim 6, further comprising a cloud management platform, wherein the cloud management platform is used for storing a plurality of pump station running state data and sewage data collected by the sensor assembly, storing and displaying pump station distribution position data and pump station state data, recording the running state of a pump station cluster, and executing pump station video retrieval, remote pump station startup and shutdown, and pump station parameter change operations;
the cloud management platform comprises the computing processing unit, and the computing processing unit is used for computing and processing the time data, the space data and the additional influence data acquired by the sensor assembly.
8. The system according to claim 6, further comprising a video monitoring system for acquiring video data of the environment surrounding the pump station, wherein the video monitoring system is connected to the cloud management platform through a video network.
9. The system according to any of claims 6-8, wherein the pump station comprises a sewage lift pump station, the sewage lift pump station comprises a control cabinet, a sewage storage tank, a water pump and the sensor assembly, the control cabinet is electrically connected with the water pump and the sensor assembly, the sewage storage tank is connected with the water pump, and the sensor assembly is fixedly connected with the sewage storage tank.
10. A storage medium, characterized in that the storage medium comprises a stored program, wherein the program, when executed, controls an apparatus in which the storage medium is located to perform the method of any of claims 1-5.
CN202010516998.6A 2020-06-09 2020-06-09 Pump station cluster intelligent prediction method and system for urban sewage treatment and storage medium Active CN111624874B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010516998.6A CN111624874B (en) 2020-06-09 2020-06-09 Pump station cluster intelligent prediction method and system for urban sewage treatment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010516998.6A CN111624874B (en) 2020-06-09 2020-06-09 Pump station cluster intelligent prediction method and system for urban sewage treatment and storage medium

Publications (2)

Publication Number Publication Date
CN111624874A true CN111624874A (en) 2020-09-04
CN111624874B CN111624874B (en) 2022-12-06

Family

ID=72271986

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010516998.6A Active CN111624874B (en) 2020-06-09 2020-06-09 Pump station cluster intelligent prediction method and system for urban sewage treatment and storage medium

Country Status (1)

Country Link
CN (1) CN111624874B (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114017302A (en) * 2021-11-03 2022-02-08 江苏四联自动化科技有限公司 Dynamic constant liquid level intelligent adjusting method and system for multi-stage pump station
CN117217503A (en) * 2023-11-09 2023-12-12 山东华宝隆轻工机械有限公司 Intelligent pump station pump group remote intelligent scheduling management system based on big data

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170262995A1 (en) * 2016-03-11 2017-09-14 Qualcomm Incorporated Video analysis with convolutional attention recurrent neural networks
CN107529651A (en) * 2017-08-18 2018-01-02 北京航空航天大学 A kind of urban transportation passenger flow forecasting and equipment based on deep learning
CN109948863A (en) * 2019-03-28 2019-06-28 杭州电子科技大学 Drainage pipeline networks inspection shaft liquid level prediction technique based on shot and long term memory models LSTM
CN110909926A (en) * 2019-11-18 2020-03-24 中国计量大学 TCN-LSTM-based solar photovoltaic power generation prediction method
CN111222698A (en) * 2020-01-06 2020-06-02 重庆邮电大学 Long-and-short-term memory network-based ponding water level prediction method for Internet of things
CN111242344A (en) * 2019-12-11 2020-06-05 大连海事大学 Intelligent water level prediction method based on cyclic neural network and convolutional neural network

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170262995A1 (en) * 2016-03-11 2017-09-14 Qualcomm Incorporated Video analysis with convolutional attention recurrent neural networks
CN107529651A (en) * 2017-08-18 2018-01-02 北京航空航天大学 A kind of urban transportation passenger flow forecasting and equipment based on deep learning
CN109948863A (en) * 2019-03-28 2019-06-28 杭州电子科技大学 Drainage pipeline networks inspection shaft liquid level prediction technique based on shot and long term memory models LSTM
CN110909926A (en) * 2019-11-18 2020-03-24 中国计量大学 TCN-LSTM-based solar photovoltaic power generation prediction method
CN111242344A (en) * 2019-12-11 2020-06-05 大连海事大学 Intelligent water level prediction method based on cyclic neural network and convolutional neural network
CN111222698A (en) * 2020-01-06 2020-06-02 重庆邮电大学 Long-and-short-term memory network-based ponding water level prediction method for Internet of things

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
KAIHAO ZHANG等: "Facial Expression Recognition Based on Deep Evolutional Spatial-Temporal Networks", 《IEEE TRANSACTIONS ON IMAGE PROCESSING》 *
MINGYANG PAN等: "Water Level Prediction Model Based on GRU and CNN", 《IEEE ACCESS》 *
薛佳瑶等: "基于卷积循环神经网络的城市区域车流量预测模型", 《信息工程大学学报》 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114017302A (en) * 2021-11-03 2022-02-08 江苏四联自动化科技有限公司 Dynamic constant liquid level intelligent adjusting method and system for multi-stage pump station
CN117217503A (en) * 2023-11-09 2023-12-12 山东华宝隆轻工机械有限公司 Intelligent pump station pump group remote intelligent scheduling management system based on big data
CN117217503B (en) * 2023-11-09 2024-02-06 山东华宝隆轻工机械有限公司 Intelligent pump station pump group remote intelligent scheduling management system based on big data

Also Published As

Publication number Publication date
CN111624874B (en) 2022-12-06

Similar Documents

Publication Publication Date Title
Gasparin et al. Deep learning for time series forecasting: The electric load case
Hafeez et al. Electric load forecasting based on deep learning and optimized by heuristic algorithm in smart grid
Dai et al. Deeptrend: A deep hierarchical neural network for traffic flow prediction
Catalão et al. Short-term wind power forecasting in Portugal by neural networks and wavelet transform
Kusiak et al. Modeling and analysis of pumps in a wastewater treatment plant: A data-mining approach
CN112288595A (en) Power grid load prediction method, related device, equipment and storage medium
CN111624874B (en) Pump station cluster intelligent prediction method and system for urban sewage treatment and storage medium
Woo et al. Predicting wind turbine power and load outputs by multi-task convolutional LSTM model
Cárdenas et al. Evolutive ANFIS training for energy load profile forecast for an IEMS in an automated factory
Shaikh et al. A new approach to seasonal energy consumption forecasting using temporal convolutional networks
Porteiro et al. Short term load forecasting of industrial electricity using machine learning
Wang et al. Improving economic values of day‐ahead load forecasts to real‐time power system operations
CN116957698A (en) Electricity price prediction method based on improved time sequence mode attention mechanism
CN117556332B (en) Heating furnace waste heat utilization management method, system and storage medium based on Internet of things
WO2019227273A1 (en) Hierarchical concept based neural network model for data center power usage effectiveness prediction
Tsai Dynamic grey platform for efficient forecasting management
CN117674128A (en) Automatic fault removal method based on power dispatching system
Peng et al. Stacked autoencoders and extreme learning machine based hybrid model for electrical load prediction
CN116578858A (en) Air compressor fault prediction and health degree evaluation method and system based on graphic neural network
CN115576203A (en) Embedded electric precipitation intelligent control method and system based on neural network
CN105652658B (en) Real power control system based on the analysis of photovoltaic DC field data and real time environment monitoring
CN116027829A (en) Machine room temperature control method, device, equipment and storage medium
Chen et al. Research on load forecasting method of distribution transformer based on deep learning
Li et al. A Novel Short-term Load Forecasting Model by TCN-LSTM Structure with Attention Mechanism
Zjavka Power Quality Estimations for Unknown Binary Combinations of Electrical Appliances Based on the Step-by-Step Increasing Model Complexity

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