CN117075507A - Sludge drying control system and method based on AI model water content online detection - Google Patents
Sludge drying control system and method based on AI model water content online detection Download PDFInfo
- Publication number
- CN117075507A CN117075507A CN202311028834.9A CN202311028834A CN117075507A CN 117075507 A CN117075507 A CN 117075507A CN 202311028834 A CN202311028834 A CN 202311028834A CN 117075507 A CN117075507 A CN 117075507A
- Authority
- CN
- China
- Prior art keywords
- data
- water content
- sludge
- model
- window
- 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.)
- Pending
Links
- 239000010802 sludge Substances 0.000 title claims abstract description 133
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 title claims abstract description 126
- 238000001514 detection method Methods 0.000 title claims abstract description 99
- 238000001035 drying Methods 0.000 title claims abstract description 58
- 238000000034 method Methods 0.000 title claims abstract description 42
- 230000007613 environmental effect Effects 0.000 claims abstract description 54
- 238000012549 training Methods 0.000 claims abstract description 36
- 238000012545 processing Methods 0.000 claims abstract description 26
- 230000008569 process Effects 0.000 claims abstract description 19
- 238000007781 pre-processing Methods 0.000 claims abstract description 15
- 238000013519 translation Methods 0.000 claims abstract description 12
- 238000012360 testing method Methods 0.000 claims description 9
- 230000004913 activation Effects 0.000 claims description 8
- 238000007599 discharging Methods 0.000 claims description 8
- 238000004140 cleaning Methods 0.000 claims description 7
- 238000011156 evaluation Methods 0.000 claims description 6
- 230000009467 reduction Effects 0.000 claims description 5
- 238000013528 artificial neural network Methods 0.000 claims description 4
- 238000002790 cross-validation Methods 0.000 claims description 4
- 238000012795 verification Methods 0.000 claims description 3
- 239000000523 sample Substances 0.000 description 10
- 230000006870 function Effects 0.000 description 9
- 230000001276 controlling effect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000013461 design Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 230000001105 regulatory effect Effects 0.000 description 3
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 2
- 230000008901 benefit Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000002759 z-score normalization Methods 0.000 description 2
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Classifications
-
- C—CHEMISTRY; METALLURGY
- C02—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F—TREATMENT OF WATER, WASTE WATER, SEWAGE, OR SLUDGE
- C02F11/00—Treatment of sludge; Devices therefor
- C02F11/12—Treatment of sludge; Devices therefor by de-watering, drying or thickening
- C02F11/121—Treatment of sludge; Devices therefor by de-watering, drying or thickening by mechanical de-watering
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01D—MEASURING NOT SPECIALLY ADAPTED FOR A SPECIFIC VARIABLE; ARRANGEMENTS FOR MEASURING TWO OR MORE VARIABLES NOT COVERED IN A SINGLE OTHER SUBCLASS; TARIFF METERING APPARATUS; MEASURING OR TESTING NOT OTHERWISE PROVIDED FOR
- G01D21/00—Measuring or testing not otherwise provided for
- G01D21/02—Measuring two or more variables by means not covered by a single other subclass
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/24—Earth materials
- G01N33/246—Earth materials for water content
-
- 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
- G05B19/00—Programme-control systems
- G05B19/02—Programme-control systems electric
- G05B19/04—Programme control other than numerical control, i.e. in sequence controllers or logic controllers
- G05B19/042—Programme control other than numerical control, i.e. in sequence controllers or logic controllers using digital processors
- G05B19/0423—Input/output
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
- G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
- G05B2219/00—Program-control systems
- G05B2219/20—Pc systems
- G05B2219/25—Pc structure of the system
- G05B2219/25257—Microcontroller
Abstract
The disclosure relates to the technical field of sludge treatment, and provides a sludge drying control system and method based on-line detection of water content of an AI model. Firstly, carrying out serialization processing on environmental parameter data, converting the environmental parameter data into time sequence data, and simultaneously carrying out sequence translation in the data preprocessing step, wherein the last data of a window is used as a label, the model training process does not need marking, and training can be carried out by directly using historical data, so that the training efficiency is improved; when the AI detection model is used for measuring the water content of the sludge, the water content of the sludge is estimated through environmental parameter data, so that the cost problem caused by purchasing a professional water content detection instrument is avoided.
Description
Technical Field
The disclosure relates to the technical field related to sludge drying control, in particular to a sludge drying control system and method based on-line detection of water content of an AI model.
Background
The statements in this section merely provide background information related to the present disclosure and may not necessarily constitute prior art.
The sludge has complex components and difficult pollution control, so the sludge is required to be dried, reduced, finally harmless and recycled, and in order to achieve the drying effect of the sludge, the water content of the sludge is required to be detected on line in the drying process, so that the dry sludge which is stable in process and meets the water content requirement is obtained.
The inventor finds that the existing sludge drying control system is not high in intellectualization, and a worker is required to check the state of sludge in the treatment process, and the running parameters of drying equipment are adjusted through experience; the special sludge moisture content detection device is adopted for detecting the sludge, the equipment is high in price, the detection head of the device needs to be stretched into the sludge, the detection device is easy to damage in a severe environment such as the sludge, and the measurement cost is increased due to the large loss of the detection device; and through the sludge detection device, the data of the sensing measurement has time lag, and the fact that the water content deviates from the design value is detected, and the readjustment equipment outputs a large amount of unqualified sludge is detected.
Disclosure of Invention
In order to solve the problems, the disclosure provides a sludge drying control system and method based on AI model water content online detection, which can be used for online detection of the water content of sludge by acquiring environmental parameter data of dried sludge, so that expensive professional detection instruments can be replaced, the real-time property of the detection of the water content of discharged sludge is ensured, and the sludge drying stability is improved.
In order to achieve the above purpose, the present disclosure adopts the following technical scheme:
one or more embodiments provide a sludge drying control system based on-line detection of the water content of an AI model, which comprises a dry sludge environment parameter acquisition device, an AI detection model and a sludge drying control module;
the dry sludge environment parameter acquisition device is used for acquiring environment parameter data of the dry sludge;
the AI detection model is configured to carry out serialization processing on the collected environmental parameter data, and calculate the water content data according to the time sequence data of the environmental parameter data at the current moment;
the sludge drying control module is configured to generate control parameter data of drying process equipment according to the calculated water content data and control the drying process in real time.
In the training process of the AI detection model, the acquired historical environmental parameter data and the corresponding water content are subjected to sequence processing and translation, the water content is used as the last data of the window, the last data of the window is used as a label to be output as the AI detection model, and other data except the label data in the window are used as input to train the AI detection model.
A sludge drying control method of a sludge drying control system based on AI model water content online detection comprises the following steps:
acquiring environmental parameter data of the dry sludge;
carrying out serialization processing on the acquired environmental parameter data, adopting a trained AI detection model, and obtaining water content data by adopting time sequence data speculation of the environmental parameter data at the current moment;
generating control parameter data of drying process equipment according to the presumed water content data, and controlling a drying process;
in the training process of the AI detection model, the acquired historical environmental parameter data and the corresponding water content are subjected to sequence processing and translation, the water content is used as the last data of the window, the last data of the window is used as a label to be output as the AI detection model, and other data except the label data in the window are used as input to train the AI detection model.
Compared with the prior art, the beneficial effects of the present disclosure are:
in the method, firstly, the environmental parameter data is subjected to serialization processing, the environmental parameter data can be converted into time sequence data, meanwhile, a sequence translation step is arranged in the data preprocessing step, the last data of a window is used as a label, the model training process does not need marking, the training can be performed by directly using historical data, and the training efficiency is improved; when the AI detection model is used for measuring the water content of the sludge, the water content of the sludge is estimated through environmental parameter data, so that the cost problem caused by purchasing a professional water content detection instrument is avoided.
The advantages of the present disclosure, as well as those of additional aspects, will be described in detail in the following detailed description of embodiments.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate and explain the exemplary embodiments of the disclosure and together with the description serve to explain and do not limit the disclosure.
FIG. 1 is a schematic diagram of a system configuration of embodiment 1 of the present disclosure;
FIG. 2 is a schematic diagram of a conventional water cut detection using a water cut detector;
FIG. 3 is a schematic diagram of data acquisition of a dry sludge environmental parameter acquisition device of embodiment 1 of the present disclosure;
fig. 4 is a flowchart of a control method of embodiment 2 of the present disclosure.
Detailed Description
The disclosure is further described below with reference to the drawings and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments in accordance with the present disclosure. As used herein, the singular is also intended to include the plural unless the context clearly indicates otherwise, and furthermore, it is to be understood that the terms "comprises" and/or "comprising" when used in this specification are taken to specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof. It should be noted that, without conflict, the various embodiments and features of the embodiments in the present disclosure may be combined with each other. The embodiments will be described in detail below with reference to the accompanying drawings.
Example 1
In the technical scheme disclosed in one or more embodiments, as shown in fig. 1 to 3, a sludge drying control system based on-line detection of the water content of an AI model comprises a dry sludge environment parameter acquisition device, an AI detection model and a sludge drying control module;
the dry sludge environment parameter acquisition device is used for acquiring environment parameter data of the dry sludge;
the AI detection model is configured to perform serialization processing on the collected environmental parameter data, and presume the water content at the moment according to the time sequence data of the environmental parameter data at the moment;
the sludge drying control module is configured to generate control parameter data of drying process equipment according to the predicted water content data and control the drying process.
In the training process of the AI detection model, the acquired historical environmental parameter data and the corresponding water content are subjected to sequence processing and translation, the water content is used as the last data of the window, the last data of the window is used as a label to be output as the AI detection model, and other data except the label data in the window are used as input to train the AI detection model.
In the embodiment, firstly, the environmental parameter data is subjected to serialization processing, the environmental parameter data can be converted into time series data, meanwhile, in the step of data preprocessing, a step of sequence translation is provided, the last data of a window is used as a label, the model training process does not need marking, the training can be performed by directly using historical data, and the training efficiency is improved; when the AI detection model is used for measuring the water content of the sludge, the water content of the sludge is estimated through environmental parameter data, so that the cost problem caused by purchasing a professional water content detection instrument is avoided. The AI detection model has high recognition speed and can realize timeliness of sludge drying control.
Meanwhile, the collected data are environmental parameter data, the test environment is not directly contacted with sludge, and damage of the measuring sensor is reduced. As shown in fig. 2 and 3, fig. 2 is a schematic diagram of detecting the water content by using a water content detector in the prior art, and the sensor in this embodiment only needs to collect environmental data, as shown in fig. 3, and does not need to be in direct contact with sludge. In addition, the sensor used for collecting the environmental data is a conventional sensor, and compared with a special sensor for the water content, the sensor has obvious price advantage, and the detection cost is greatly reduced.
The sludge drying control system based on the on-line detection of the sludge moisture content of the AI model can be realized by combining software and hardware, and is integrated in the sludge drying control system equipment of the on-line detection of the sludge moisture content of the AI model. The sludge drying control system equipment for the on-line detection of the AI sludge water content can be equipment such as a computer.
Optionally, as shown in fig. 1, the sludge drying control system further comprises a feeding device, a sludge dryer and a discharging device.
The feeding equipment is used for conveying sludge to be dried to the sludge dryer; in particular, the feeding device may comprise a wet sludge silo, and a wet sludge transfer system.
The sludge dryer is used for conveying dried sludge to the discharging equipment;
and the discharging equipment is used for conveying or storing the dried sludge. Specifically, the discharging equipment comprises a dry sludge conveying system and a dry sludge storage bin.
In this embodiment, the device for collecting environmental parameters of dry sludge is disposed at the outlet end of the sludge drier, and specifically disposed on the pipe wall of the dry sludge conveying system. And collecting the temperature and humidity of the gas released by the dry sludge in the discharging equipment, and transmitting the temperature, the humidity and the like to a client, wherein the client is internally provided with the water content AI detection model for realizing the detection of the water content of the dry sludge. The water content AI detection model transmits the calculated water content to a sludge drying control module, and the sludge drying control module transmits control parameters to the feeding equipment and the sludge dryer. And an internal logic control sends out an adjusting instruction to change the operation parameters of executing mechanisms such as a wet sludge conveying system, a sludge dryer and the like, and finally, the water content of the dried sludge is ensured to be maintained within a design range.
Optionally, the method for controlling the drying process equipment by the sludge drying control module after receiving the water content data comprises the following steps:
a1, when the water content is higher than a set value, performing a first water content reduction regulation operation; otherwise, performing the water content increasing regulation and control operation;
the first water content reduction regulation operation can comprise the steps of increasing the height of a weir plate of the sludge drier, reducing the rotating speed of a shaft and the like; the water content can be reduced by means of increasing the height of the weir plate, reducing the shaft rotation speed, and the like. The increasing water content regulating operation may include: the water content is improved by reducing the height of a weir plate of the sludge dryer and the rotating speed of a lifting shaft.
A2, acquiring a water content detection value in a time interval set after the first water content reduction regulation and control operation, wherein the water content still continuously rises, and further adopting a second operation, wherein the second operation can be used for reducing the conveying amount of wet sludge; the water content can be further controlled by reducing the conveying amount of wet sludge;
and A3, acquiring a water content detection value, and executing the step 1. Steps 1 to 3 are cyclically performed so that the water content is within a set range.
In the embodiment, when the water content is regulated and controlled, the sectional operation is performed, so that the fineness of regulating and controlling the water content can be improved, the sudden rise and the sudden fall are avoided, and the accuracy of sludge treatment regulation and control is improved.
In the embodiment, after the water content value of the dried sludge is obtained through calculation, through written internal logic, the operation parameters of equipment such as a wet sludge conveying system, a sludge drier and the like are adjusted, and finally, the water content of the dried sludge is kept within a design range.
As shown in fig. 1, in one implementation manner, the device for collecting the environmental parameters of the dried sludge is arranged in the dry sludge conveying system and is used for collecting the environmental parameters of the dried sludge.
Optionally, the dry sludge environmental parameter acquisition device comprises probes such as a temperature sensor, a humidity sensor and the like, and is not a professional probe for directly measuring the water content of the sludge.
In some embodiments, the AI detection model may be constructed using a long-term or short-term memory network, and includes an input layer, a hidden layer, an activation function layer, and an output layer connected in sequence, where the input layer is an LSTM layer, the hidden layer includes a plurality of LSTM layers, the activation function layer selects a relu function, and the output layer is a Dense layer.
The Dense layer is a fully connected neural network layer and is used for extracting the correlation between the features after nonlinear change of the Dense and finally mapping the correlation to an output space.
The embodiment is provided with the hidden layer of the stacked LSTM, can increase the network depth, enhance the fitting capacity, approximate the highly nonlinear function by using fewer parameters, and improve the accuracy of water content prediction.
Further, the system also comprises a preprocessing module, wherein the preprocessing module is used for preprocessing data before the AI detection model is adopted to predict the water content, and comprises the following preprocessing processes:
1. performing data cleaning on the acquired environmental parameter data; and identifying whether the acquired environmental parameter data has a missing value, namely whether the temperature and humidity data has the missing value, and deleting the row where the missing value is located if the missing value exists.
2. Carrying out standardization treatment on the cleaned data;
optionally, for temperature and humidity data, the data is normalized by using a Z-score normalization method, so that the data is distributed in a smaller range.
3. And carrying out serialization processing on the standardized data, converting the original temperature and humidity data into time series data, and dividing the data into a plurality of time windows by adopting a sliding window method, wherein each window contains N data points.
After pretreatment, the data of each data window are input into an AI detection model to predict the water content.
4. In the AI detection model training stage, the training comprises the steps of acquiring environmental parameter data and corresponding water content data, taking the water content data as the last data point in the serialization processing of the historical data preprocessing stage, translating the time sequence data obtained by serialization, taking the water content data of the last data point of each time window as a label, and translating the data point of the corresponding time window forward by one bit. The prediction can be rolled simultaneously in the prediction process, and the accuracy of the water content prediction can be improved.
Further, the method for training the AI detection model comprises the following steps:
s1, acquiring environmental parameter data of dry sludge and corresponding water content of the dry sludge, and constructing a data sample set;
step S2, dividing the data set: dividing the data set into a training set, a verification set and a test set;
step S3, preprocessing the divided data sample set data, respectively carrying out data cleaning and standardization processing, carrying out translation on window data obtained by serialization after dividing the window data, taking the last data point of each time window as a label, and translating the data point of the corresponding time window forward by one bit;
specifically, the tag data is used as a tag of the previous window and also used as the first data of the next window, the translated data is data of (t-n) to (t) sections based on the current moment, the data of (t-n) to (t-1) are input data, and the moisture content of the t moment is tag data.
S4, taking the data of the preprocessed data time window as input, taking the water content of one data point after the time window as output, and inputting the data into an AI detection model for training;
s5, minimizing the prediction error of the model based on a gradient descent algorithm, and simultaneously performing cross validation and parameter adjustment until the prediction accuracy is met, so as to obtain a trained AI detection model;
and S6, evaluating the performance of the trained AI detection model by using a test set, and adjusting and optimizing the model according to an evaluation result.
Alternatively, during training, the loss function may use the loss of L1, with the formula:
wherein mae represents an average value of absolute errors between the predicted value of the water content and the detected actual value; y is i Andthe water content actual value and the corresponding estimated value of the ith sample are respectively represented, the corresponding estimated value is the value obtained by the AI detection model, and N is the number of samples.
In step S6, the adjusting and optimizing optimizer selects Adam optimizer, and the evaluation index selects root mean square error, where the formula is:
wherein y is i Andthe actual value and the corresponding estimated value of the water content of the ith sample are respectively shown, and N is the number of samples.
Example 2
Based on embodiment 1, in this embodiment, a sludge drying control method based on the sludge drying control system based on the online detection of the water content of the AI model described in embodiment 1 is provided, as shown in fig. 4, and includes the following steps:
step 1, acquiring environmental parameter data of dry sludge;
step 2, carrying out serialization processing on the acquired environmental parameter data, adopting a trained AI detection model, and obtaining water content data by adopting time sequence data of the environmental parameter data at the current moment in a presumption mode;
step 3, generating control parameter data of the drying process equipment according to the presumed water content data, and controlling the drying process;
in the training process of the AI detection model, the acquired historical environmental parameter data and the corresponding water content are subjected to sequence translation, the last data of the window, namely the water content, is used as a label to be output to the AI detection model, other data except the label data in the window is used as input, and the AI detection model is trained.
In the embodiment, firstly, the environmental parameter data is subjected to serialization processing, the environmental parameter data can be converted into time series data, meanwhile, in the step of data preprocessing, a step of sequence translation is provided, the last data of a window is used as a label, the model training process does not need marking, the training can be performed by directly using historical data, and the training efficiency is improved; when the AI detection model is used for measuring the water content of the sludge, the water content of the sludge is estimated through environmental parameter data, so that the cost problem caused by purchasing a professional water content detection instrument is avoided.
Optionally, the AI detection model includes an input layer, a hidden layer, an activation function layer and an output layer which are sequentially connected, the input layer is an LSTM layer, the hidden layer includes a plurality of LSTM layers, the activation function layer selects a relu function, and the output layer is a fully connected neural network layer.
Before the water content is presumed, the data needs to be preprocessed, which comprises the following preprocessing processes:
1. performing data cleaning on the acquired environmental parameter data; and identifying whether the acquired environmental parameter data has a missing value, namely whether the temperature and humidity data has the missing value, and deleting the row where the missing value is located if the missing value exists.
2. Carrying out standardization treatment on the cleaned data;
optionally, for temperature and humidity data, the data is normalized by using a Z-score normalization method, so that the data is distributed in a smaller range.
3. And carrying out serialization processing on the standardized data, converting the original temperature and humidity data into time series data, and dividing the data into a plurality of time windows by adopting a sliding window method, wherein each window contains N data points.
After pretreatment, the data of each data window are input into an AI detection model to predict the water content.
4. In the AI detection model training stage, the training comprises the steps of acquiring environmental parameter data and corresponding water content data, taking the water content data as the last data point in the serialization processing of the historical data preprocessing stage, translating the time sequence data obtained by serialization, taking the water content data of the last data point of each time window as a label, and translating the data point of the corresponding time window forward by one bit. The prediction can be rolled simultaneously in the prediction process, and the accuracy of the water content prediction can be improved.
Further, the method for training the AI detection model comprises the following steps:
s1, acquiring environmental parameter data of dry sludge and corresponding water content of the dry sludge, and constructing a data sample set;
step S2, dividing the data set: the dataset was divided into three parts, training set, validation set and test set, with proportions of 70%, 15% and 15%.
Step S3, preprocessing the divided data sample set data, respectively carrying out data cleaning and standardization processing, carrying out translation on window data obtained by serialization after dividing the window data, taking the last data point of each time window as a label, and translating the data point of the corresponding time window forward by one bit;
specifically, the tag data is used as a tag of the previous window and also used as the first data of the next window, the translated data is data of (t-n) to (t) sections based on the current moment, the data of (t-n) to (t-1) are input data, and the moisture content of the t moment is tag data.
S4, taking the data of the preprocessed data time window as input, taking the water content of one data point after the time window as output, and inputting the data into an AI detection model for training;
s5, minimizing the prediction error of the model based on a gradient descent algorithm, and simultaneously performing cross validation and parameter adjustment until the prediction accuracy is met, so as to obtain a trained AI detection model;
and S6, evaluating the performance of the trained AI detection model by using a test set, and adjusting and optimizing the model according to an evaluation result.
Alternatively, in the training process, the loss function may use the loss of L1, where the formula is:
wherein mae represents an average value of absolute errors between the predicted value of the water content and the detected actual value; y is i Andthe water content actual value and the corresponding estimated value of the ith sample are respectively represented, the corresponding estimated value is the value obtained by the AI detection model, and N is the number of samples.
In step S6, the adjusting and optimizing optimizer selects Adam optimizer, and the evaluation index selects root mean square error, where the formula is:
wherein y is i Andthe actual value and the corresponding estimated value of the water content of the ith sample are respectively shown, and N is the number of samples.
The foregoing description of the preferred embodiments of the present disclosure is provided only and not intended to limit the disclosure so that various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.
While the specific embodiments of the present disclosure have been described above with reference to the drawings, it should be understood that the present disclosure is not limited to the embodiments, and that various modifications and changes can be made by one skilled in the art without inventive effort on the basis of the technical solutions of the present disclosure while remaining within the scope of the present disclosure.
Claims (10)
1. Sludge drying control system based on AI model moisture content on-line measuring, its characterized in that: the system comprises a dry sludge environment parameter acquisition device, an AI detection model and a sludge drying control module;
the dry sludge environment parameter acquisition device is used for acquiring environment parameter data of the dry sludge;
the AI detection model is configured to carry out serialization processing on the collected environmental parameter data, and calculate the water content data according to the time sequence data of the environmental parameter data at the current moment;
the sludge drying control module is configured to generate control parameter data of drying process equipment according to the calculated water content data and control the drying process in real time;
in the training process of the AI detection model, the acquired historical environmental parameter data and the corresponding water content are subjected to sequence processing and translation, the water content is used as the last data of the window, the last data of the window is used as a label to be output as the AI detection model, and other data except the label data in the window are used as input to train the AI detection model.
2. The sludge drying control system based on the online detection of the water content of the AI model as claimed in claim 1, wherein: the dry sludge environment parameter acquisition device comprises a temperature sensor and a humidity sensor.
3. The sludge drying control system based on the online detection of the water content of the AI model as claimed in claim 1, wherein: the AI detection model comprises an input layer, a hidden layer, an activation function layer and an output layer which are sequentially connected, wherein the input layer is an LSTM layer, the hidden layer comprises a plurality of LSTM layers, the activation function layer selects a relu function, and the output layer is a fully-connected neural network layer.
4. The sludge drying control system based on the on-line detection of the water content of the AI model as claimed in claim 1, wherein the pretreatment of the collected data comprises the following processes:
performing data cleaning on the acquired environmental parameter data;
carrying out standardization treatment on the cleaned data;
and carrying out serialization processing on the standardized data, converting the original temperature and humidity data into time series data, and dividing the data into a plurality of time windows by adopting a sliding window method, wherein each window contains N data points.
5. The sludge drying control system based on the online detection of the water content of the AI model as claimed in claim 1, further comprising a method for training the AI detection model, comprising the steps of:
acquiring environmental parameter data of the dry sludge and the corresponding water content of the dry sludge, and constructing a data sample set;
dividing the data set into a training set, a verification set and a test set;
preprocessing the divided data sample set data, respectively performing data cleaning and standardization processing, and dividing window data after serialization processing, wherein in the serialization process, the water content is used as the last data point at the corresponding moment; translating the window data obtained by serialization, taking the last data point of each time window as a label, and translating the data point of the corresponding time window forward by one bit;
taking the data of the preprocessed data time window as input, taking the water content of the last data point of the time window as output, and inputting the data into an AI detection model for training;
minimizing the prediction error of the model based on a gradient descent algorithm, and simultaneously performing cross validation and parameter adjustment until the prediction accuracy is met, so as to obtain a trained AI detection model;
and evaluating the performance of the trained AI detection model by using the test set, and adjusting and optimizing the model according to the evaluation result.
6. The sludge drying control system based on the online detection of the water content of the AI model as claimed in claim 1, wherein: the sludge drying control system also comprises feeding equipment, a sludge dryer and discharging equipment;
the feeding equipment is used for conveying sludge to be dried to the sludge dryer;
the sludge dryer is used for conveying dried sludge to the discharging equipment;
the discharging device is used for conveying or storing the dried sludge;
the dry sludge environmental parameter acquisition device is arranged at the outlet end of the sludge drier.
7. The sludge drying control system based on the online detection of the water content of the AI model as claimed in claim 1, wherein:
the method for controlling the drying process equipment by the sludge drying control module after receiving the water content data comprises the following steps:
step 1, when the water content is higher than a set range, performing a first water content reduction regulation operation; otherwise, when the water content is lower than the set range, performing the water content increasing regulation and control operation;
step 2, acquiring a water content detection value in a time interval set after the first water content reduction regulation operation, wherein the water content still continuously rises, and further adopting a second operation;
and 3, acquiring a water content detection value, executing the step 1, and circularly executing the steps 1 to 3 so that the water content is within a set range.
8. The sludge drying control method based on the sludge drying control system based on the on-line detection of the water content of the AI model as claimed in any one of claims 1 to 7, characterized by comprising the following steps:
acquiring environmental parameter data of the dry sludge;
carrying out serialization processing on the acquired environmental parameter data, adopting a trained AI detection model, and obtaining water content data by adopting time sequence data speculation of the environmental parameter data at the current moment;
generating control parameter data of drying process equipment according to the presumed water content data, and controlling a drying process;
in the training process of the AI detection model, the acquired historical environmental parameter data and the corresponding water content are subjected to sequence processing and translation, the water content is used as the last data of the window, the last data of the window is used as a label to be output as the AI detection model, and other data except the label data in the window are used as input to train the AI detection model.
9. The sludge drying control method as claimed in claim 8, wherein:
the AI detection model comprises an input layer, a hidden layer, an activation function layer and an output layer which are sequentially connected, wherein the input layer is an LSTM layer, the hidden layer comprises a plurality of LSTM layers, the activation function layer selects a relu function, and the output layer is a fully-connected neural network layer.
10. The sludge drying control method as claimed in claim 8, wherein: a method of training an AI detection model, comprising the steps of:
acquiring environmental parameter data of the dry sludge and the corresponding water content of the dry sludge, and constructing a data sample set;
dividing the data set into a training set, a verification set and a test set;
preprocessing the divided data sample set data, respectively carrying out data cleaning and standardization processing, carrying out serialization processing, dividing window data, translating the window data obtained by serialization, taking the last data point of each time window, namely the water content, as a label, and translating the data point of the corresponding time window forwards by one bit;
taking the data of the preprocessed data time window as input, taking the water content of one data point after the time window as output, and inputting the data into an AI detection model for training;
minimizing the prediction error of the model based on a gradient descent algorithm, and simultaneously performing cross validation and parameter adjustment until the prediction accuracy is met, so as to obtain a trained AI detection model;
and evaluating the performance of the trained AI detection model by using the test set, and adjusting and optimizing the model according to the evaluation result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311028834.9A CN117075507A (en) | 2023-08-15 | 2023-08-15 | Sludge drying control system and method based on AI model water content online detection |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311028834.9A CN117075507A (en) | 2023-08-15 | 2023-08-15 | Sludge drying control system and method based on AI model water content online detection |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117075507A true CN117075507A (en) | 2023-11-17 |
Family
ID=88709072
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311028834.9A Pending CN117075507A (en) | 2023-08-15 | 2023-08-15 | Sludge drying control system and method based on AI model water content online detection |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117075507A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117519042A (en) * | 2023-11-30 | 2024-02-06 | 天瑞集团信息科技有限公司 | Intelligent control method, system and storage medium for cement production based on AI technology |
-
2023
- 2023-08-15 CN CN202311028834.9A patent/CN117075507A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117519042A (en) * | 2023-11-30 | 2024-02-06 | 天瑞集团信息科技有限公司 | Intelligent control method, system and storage medium for cement production based on AI technology |
| CN117519042B (en) * | 2023-11-30 | 2024-04-26 | 天瑞集团信息科技有限公司 | Intelligent control method, system and storage medium for cement production based on AI technology |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111275288B (en) | XGBoost-based multidimensional data anomaly detection method and device | |
| Alhoniemi et al. | Process monitoring and modeling using the self-organizing map | |
| CN116502169B (en) | Centrifugal dehydrator working state detection method based on data detection | |
| Hollmén | User profiling and classification for fraud detection in mobile communications networks | |
| CN111275335A (en) | Data-driven slurry circulating pump optimization method and system | |
| CN117075507A (en) | Sludge drying control system and method based on AI model water content online detection | |
| CN109034191B (en) | ELM-based one-dimensional telemetry data abnormal interpretation method | |
| CN111562108A (en) | Rolling bearing intelligent fault diagnosis method based on CNN and FCMC | |
| CN110650058B (en) | Network traffic analysis method, device, storage medium and equipment | |
| CN109623489B (en) | Improved machine tool health state evaluation method and numerical control machine tool | |
| CN108090629B (en) | Load prediction method and system based on nonlinear autoregressive neural network | |
| CN117196353B (en) | Environmental pollution assessment and monitoring method and system based on big data | |
| WO2021074990A1 (en) | Search device, search method, search program, and learning model search system | |
| CN116643949A (en) | Multi-model edge cloud load prediction method and device based on VaDE clustering | |
| CN115824993A (en) | Method and device for determining chemical oxygen demand of water body, computer equipment and medium | |
| CN110956331A (en) | Method, system and device for predicting operation state of digital factory | |
| Fu et al. | MCA-DTCN: A novel dual-task temporal convolutional network with multi-channel attention for first prediction time detection and remaining useful life prediction | |
| CN117371608A (en) | Pig house multi-point temperature and humidity prediction method and system based on deep learning | |
| CN112541296A (en) | SO2 prediction method based on PSO-LSSVM | |
| CN116644348A (en) | Cross-mechanical part fault diagnosis method and device based on transfer type countermeasure migration | |
| CN114492636B (en) | Transformer winding state signal acquisition system | |
| CN112069621B (en) | Method for predicting residual service life of rolling bearing based on linear reliability index | |
| CN111220565B (en) | CPLS-based infrared spectrum measuring instrument calibration migration method | |
| Li et al. | Knowledge enhanced ensemble method for remaining useful life prediction under variable working conditions | |
| Wang | A new variable selection method for soft sensor based on deep learning |
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 |