CN112925271A - Data processing method, automatic control method and device - Google Patents

Data processing method, automatic control method and device Download PDF

Info

Publication number
CN112925271A
CN112925271A CN201911243442.8A CN201911243442A CN112925271A CN 112925271 A CN112925271 A CN 112925271A CN 201911243442 A CN201911243442 A CN 201911243442A CN 112925271 A CN112925271 A CN 112925271A
Authority
CN
China
Prior art keywords
variable
data
target operation
controlled variable
target
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
Application number
CN201911243442.8A
Other languages
Chinese (zh)
Inventor
郭立帆
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.)
Alibaba Group Holding Ltd
Original Assignee
Alibaba Group Holding 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 Alibaba Group Holding Ltd filed Critical Alibaba Group Holding Ltd
Priority to CN201911243442.8A priority Critical patent/CN112925271A/en
Publication of CN112925271A publication Critical patent/CN112925271A/en
Pending legal-status Critical Current

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
    • G05B19/00—Programme-control systems
    • G05B19/02—Programme-control systems electric
    • G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
    • G05B19/4183—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by data acquisition, e.g. workpiece identification
    • 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
    • G05B19/418—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM]
    • G05B19/41885—Total factory control, i.e. centrally controlling a plurality of machines, e.g. direct or distributed numerical control [DNC], flexible manufacturing systems [FMS], integrated manufacturing systems [IMS] or computer integrated manufacturing [CIM] characterised by modeling, simulation of the manufacturing system
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/02—Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Landscapes

  • Engineering & Computer Science (AREA)
  • Manufacturing & Machinery (AREA)
  • General Engineering & Computer Science (AREA)
  • Quality & Reliability (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Automation & Control Theory (AREA)
  • Feedback Control In General (AREA)

Abstract

The embodiment of the application provides a data processing method, an automatic control method and an automatic control device. Determining at least one controlled variable and at least one target operation variable; acquiring first time sequence data of at least one controlled variable and second time sequence data of at least one target operation variable from historical production data according to a first sampling interval; training a first model by utilizing respective first time sequence data of at least one controlled variable and respective second time sequence data of at least one target operation variable; the first model predicts the predicted value of at least one controlled variable at the current time step based on the operation data of at least one target operation variable at the previous time step; the predicted value of the at least one controlled variable is used for searching target operation data of the at least one operation variable at the current time step when the at least one controlled variable meets a preset condition. The technical scheme provided by the embodiment of the application improves the accuracy of the automatic control process.

Description

Data processing method, automatic control method and device
Technical Field
The embodiment of the application relates to the technical field of automatic control, in particular to a data processing method, an automatic control method and an automatic control device.
Background
Automatic Control (Automatic Control) means that a certain working state or parameter in a production machine or a production process automatically runs according to a predetermined rule by using an external Control device under the condition that no person directly participates in the Control, and the Control efficiency can be greatly improved through Automatic Control, so that the Automatic Control is widely applied to industrial production.
Wherein, the controlled variable refers to the process parameters to be controlled in the automatic control process, such as temperature, humidity, air pressure, etc.; the manipulated variable means a control object such as a heater, an air cooler, and the like, which can change a controlled variable in automatic control, and operation data of the manipulated variable, such as a heating degree of the heater, and the like, needs to be known to realize automatic control.
At present, in order to realize automatic control, a model control prediction method is often used, and a controlled variable is predicted through a prediction model based on historical operation data of an operation variable. Then, the current input value of the manipulated variable is optimized based on the predicted value and the expected value of the controlled variable, so that the obtained current input value of the manipulated variable is optimized for automatic control. In the prior art, a mechanism model is generally adopted as a prediction model, and the mechanism model is an accurate mathematical model established by combining the knowledge of artificial experience on the production process. Then, due to the fact that actual production conditions are complex, a mechanism model established manually cannot accurately express the complex relation between the controlled variable and the operation variable, and the existing automatic control process is inaccurate.
Disclosure of Invention
The embodiment of the application provides a data processing method, an automatic control method and an automatic control device, and improves the automatic control accuracy.
In a first aspect, an embodiment of the present application provides a data processing method, including:
determining at least one controlled variable and at least one target operating variable;
acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data according to a first sampling interval;
training a first model by using the respective first time-series data of the at least one controlled variable and the respective second time-series data of the at least one target operating variable;
the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step.
In a second aspect, an embodiment of the present application provides an automatic control method, including:
determining at least one controlled variable and at least one target operating variable;
predicting and obtaining a predicted value of the at least one controlled variable at the current time step by utilizing a first model based on the operation data of the at least one target operation variable at the previous time step; the first model is obtained by training based on first time-series data corresponding to the at least one controlled variable and second time-series data corresponding to the at least one target operation variable acquired from historical production data;
searching target operation data of the at least one target operation variable when the at least one controlled variable meets a preset condition based on a difference value between the predicted value and the expected value of the at least one controlled variable and a limit value of the at least one target operation variable;
and taking the target operation data of the at least one target operation variable as a control input value of the current time step for automatic control.
In a third aspect, an embodiment of the present application provides a data processing method, including:
determining at least one controlled variable and at least one target operation variable related to cement production business;
acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data according to a first sampling interval;
training a first model by using the respective first time-series data of the at least one controlled variable and the respective second time-series data of the at least one target operating variable;
the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step.
In a fourth aspect, an embodiment of the present application provides an automatic control method, including:
determining at least one controlled variable and at least one target operation variable related to cement production business;
predicting and obtaining a predicted value of the at least one controlled variable at the current time step by utilizing a first model based on the operation data of the at least one target operation variable at the previous time step; the first model is obtained by training based on first time-series data corresponding to the at least one controlled variable and second time-series data corresponding to the at least one target operation variable acquired from historical production data;
searching target operation data of the at least one target operation variable when the at least one controlled variable meets a preset condition based on a difference value between the predicted value and the expected value of the at least one controlled variable and a limit value of the at least one target operation variable;
and taking the target operation data of the at least one target operation variable as a control input value of the current time step to automatically control the cement production.
In a fifth aspect, an embodiment of the present application provides a data processing apparatus, including:
the first determining module is used for determining at least one controlled variable and at least one target operation variable;
the data acquisition module is used for acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data according to a first sampling interval;
the model training module is used for training a first model by utilizing first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable;
the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step.
In a sixth aspect, an embodiment of the present application provides an automatic control device, including:
the first determining module is used for determining at least one controlled variable and at least one target operation variable;
the prediction module is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step by utilizing a first model based on the operation data of the at least one target operation variable at the previous time step; the first model is obtained by training based on first time-series data corresponding to the at least one controlled variable and second time-series data corresponding to the at least one target operation variable acquired from historical production data;
the optimization module is used for searching target operation data of the at least one target operation variable when the at least one controlled variable meets a preset condition based on a difference value between a predicted value and an expected value of the at least one controlled variable and a limit value of the at least one target operation variable;
and the control module is used for taking the target operation data of the at least one target operation variable as a control input value of the current time step to carry out automatic control.
In a seventh aspect, an embodiment of the present application provides a computing device, including a processing component and a storage component;
the storage component stores one or more computer instructions; the one or more computer instructions to be invoked for execution by the processing component;
the processing component is to:
determining at least one controlled variable and at least one target operating variable;
acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data according to a first sampling interval;
training a first model by using the respective first time-series data of the at least one controlled variable and the respective second time-series data of the at least one target operating variable;
the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step.
In an eighth aspect, an embodiment of the present application provides a control device, including a processing component and a storage component;
the storage component stores one or more computer instructions; the one or more computer instructions to be invoked for execution by the processing component;
the processing component is to:
determining at least one controlled variable and at least one target operating variable;
predicting and obtaining a predicted value of the at least one controlled variable at the current time step by utilizing a first model based on the operation data of the at least one target operation variable at the previous time step; the first model is obtained by training based on first time-series data corresponding to the at least one controlled variable and second time-series data corresponding to the at least one target operation variable acquired from historical production data;
searching target operation data of the at least one target operation variable when the at least one controlled variable meets a preset condition based on a difference value between the predicted value and the expected value of the at least one controlled variable and a limit value of the at least one target operation variable;
and taking the target operation data of the at least one target operation variable as a control input value of the current time step for automatic control.
In the embodiment of the application, at least one controlled variable and at least one target operation variable are determined; acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical service data according to sampling intervals; training a first model by using the respective first time-series data of the at least one controlled variable and the respective second time-series data of the at least one target operating variable; the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step; the predicted value of the at least one controlled variable is used for searching target operation data of the at least one operation variable at the current time step when the at least one controlled variable meets a preset condition, and automatic control can be performed based on the target operation data. According to the embodiment of the application, the first model is adopted to represent the prediction model, and the accuracy of the prediction model is improved by utilizing time series data training, so that the accuracy of the automatic control process is improved.
These and other aspects of the present application will be more readily apparent from the following description of the embodiments.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative efforts.
FIG. 1 is a flow chart illustrating one embodiment of a data processing method provided herein;
FIG. 2 is a schematic diagram illustrating a model structure of a first model in a practical application according to an embodiment of the present application;
FIG. 3 illustrates a flow chart of one embodiment of an automated control method provided herein;
FIG. 4 is a diagram illustrating an automatic control process in one practical application of the embodiment of the present application;
FIG. 5 is a block diagram illustrating an embodiment of a data processing apparatus provided herein;
FIG. 6 illustrates a schematic structural diagram of one embodiment of a computing device provided herein;
FIG. 7 is a block diagram illustrating an embodiment of a data processing apparatus provided herein;
fig. 8 is a schematic structural diagram illustrating an embodiment of a control device provided in the present application.
Detailed Description
In order to make the technical solutions better understood by those skilled in the art, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application.
In some of the flows described in the specification and claims of this application and in the above-described figures, a number of operations are included that occur in a particular order, but it should be clearly understood that these operations may be performed out of order or in parallel as they occur herein, the number of operations, e.g., 101, 102, etc., merely being used to distinguish between various operations, and the number itself does not represent any order of performance. Additionally, the flows may include more or fewer operations, and the operations may be performed sequentially or in parallel. It should be noted that, the descriptions of "first", "second", etc. in this document are used for distinguishing different messages, devices, modules, etc., and do not represent a sequential order, nor limit the types of "first" and "second" to be different.
The technical scheme of the embodiment of the application can be applied to various automatic process control scenes, for example, the automatic control method is suitable for automatic production control in various industrial production processes, such as cement production automatic control scenes and the like.
As described in the background art, the prediction model in the prior art adopts a mechanism model, and then the mechanism model does not necessarily accord with the actual production working condition, so that the model prediction result is not accurate enough, thereby affecting the accuracy of the automatic control process and causing excessive energy consumption.
In order to improve the accuracy of an automatic control process and reduce energy consumption, the inventor researches and discovers that the operation data of an operation variable at a certain moment may not affect the controlled data of a controlled variable at the moment, but may affect the future controlled data of the controlled variable, and in an actual working condition, one controlled variable may be actually affected by a plurality of operation variables; acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical service data according to sampling intervals; training a first model by using the respective first time-series data of the at least one controlled variable and the respective second time-series data of the at least one target operating variable; the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step; the predicted value of the at least one controlled variable is used for searching target operation data of the at least one operation variable at the current time step when the at least one controlled variable meets a preset condition, and automatic control can be performed based on the target operation data. According to the embodiment of the application, the first model is adopted to represent the prediction model, and the prediction model is obtained by utilizing time series data training, so that the model prediction of multiple input and multiple output is realized, the accuracy of the prediction model is improved, and the accuracy of the automatic control process is improved.
The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, 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 application.
Fig. 1 is a flowchart of an embodiment of a data processing method provided in the embodiment of the present application, and the embodiment explains a technical solution of the present application from the perspective of model training. The method may comprise the steps of:
101: at least one controlled variable and at least one target operating variable are determined.
As an alternative, the at least one controlled variable and the at least one target manipulated variable may refer to a controlled variable and a manipulated variable involved in a production business.
In conjunction with the production business process, the controlled variables and manipulated variables involved in the production process can be determined.
Alternatively, the at least one target manipulated variable may be an effective manipulated variable determined from the at least one manipulated variable involved in the production business, since some manipulated variables may affect the controlled variable and some manipulated variables may not affect the controlled variable in the actual working condition. The determination of the effective operating variables is described in detail in the examples below.
102: and acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data according to a first sampling interval.
The historical production data may refer to historical production data generated by a production process or may refer to historical production data generated within a predetermined historical time period from a current time.
The time duration corresponding to the predetermined historical time period may refer to a time duration taken for one production process.
Therefore, in some embodiments, the obtaining the first time-series data of the at least one controlled variable and the second time-series data of the at least one target operation variable from the historical production data at the first sampling interval may include:
acquiring and obtaining first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data within a preset historical time period from the current time according to a first sampling interval; wherein the first time series data comprises operational data for N time steps and the second time series data comprises controlled data for M time steps; wherein M is an integer of 1 or more.
The first sampling interval may be set according to practical applications, for example, the predetermined historical time period is X minutes, and the first sampling interval may be 60 seconds.
The predetermined historical time period may be divided into M time steps, such as from T, for data acquisition at a first sampling interval1、T2……TM. For each controlled variable, controlled data corresponding to each time step can be acquired, and for each target operating variable, operating data corresponding to each time step can be acquired.
For example, assume that k target operating variables av are determined1、av2、av3……avkAnd n controlled variables cv1、av2、av3……cvn(ii) a Obtaining M time steps T according to sampling interval division1、T2……TmEach corresponding to a certain historical time. Then from T1、T2……TMThe first time series data of any one of the target manipulated variables may include avj1、avj2、avj3……avjMWhere j is 1, 2 … … k, the second time series data of any one of the controlled variables may include cvi1、avi2、avi3……cviMWherein, i takes the values of 1 and 2 … … n.
Optionally, the controlling the manipulated variables may be performed by a Distributed Control System (DCS) in an automatic Control System, and thus, optionally, the obtaining the first time-series data of the at least one controlled variable and the second time-series data of the at least one target manipulated variable from the historical production data at the first sampling interval includes:
and acquiring first time sequence data corresponding to the at least one controlled variable and second time sequence data corresponding to the at least one target operating variable from the historical production data of the DCS according to a first sampling interval.
103: and training the first model by using the first time sequence data of the at least one controlled variable and the second time sequence data of the at least one target operation variable.
The first model may be trained by using the first time-series data of the at least one controlled variable as model input data and the second time-series data of the at least one target manipulated variable as model output results, so that the first model may learn to obtain a relationship between the at least one target manipulated variable and the at least one controlled variable.
The first model may be implemented in a network structure that can process time series data, for example, the first model may be a deep learning model. The deep learning model may specifically select a Neural Network model, the Neural Network model has good data fitting capability, and may be used in a prediction scenario, the Neural Network model may be, for example, an RNN (Recurrent Neural Network ), and the like, the RNN may further select, for example, an LSTM (Long Short-Term Memory Network) for implementation, the LSTM is a time Recurrent Neural Network, and in addition, a model implementation that may perform time prediction, such as a GRU (gated Recurrent Neural Network), a trasnformer (transformation Network, a brand new Neural Network architecture based on the self-attention machine system), and the like, may also be selected.
The first model is used as a prediction model in model control prediction, and can predict and obtain a predicted value of the at least one controlled variable at a current time step based on operation data of the at least one target operation variable at a time step before the current time step; the predicted value of the at least one controlled variable is used for searching target operation data of the at least one target operation variable at the current time step when the at least one controlled variable meets a preset condition. The target operation data of the at least one target operation variable can be used as the control input value of the current time step for automatic control.
Optionally, the automatic control process may be that at least one target operation variable is controlled to operate according to target operation data, when the DCS system performs operation variable control, the target operation data of the at least one target operation variable at the current time step is used to be sent to the distributed control system DCS, and the DCS takes the target operation data of the at least one target operation variable as a control input value of the current time step and controls the at least one target operation variable to operate according to the control input value of the at least one target operation variable.
In the embodiment, the first model is adopted to represent the prediction model and is obtained by utilizing time series data training, so that the accuracy of the prediction model is improved, and the accuracy of the automatic control process is improved.
In some embodiments, the determining at least one controlled variable and at least one target operating variable may include:
determining at least one controlled variable and at least one operating variable involved in a production service;
determining at least one target manipulated variable related to the at least one controlled variable from the at least one manipulated variable.
Alternatively, a target manipulated variable associated with each controlled variable, each possibly corresponding to at least one target manipulated variable, may be determined.
In some embodiments, the determining at least one target manipulated variable related to the at least one controlled variable from the at least one manipulated variable may include:
acquiring third time series data of the controlled variables and fourth time series data of the operation variables according to a second sampling interval for each controlled variable and each operation variable;
checking whether the correlation exists between the operation variable and the controlled variable or not by utilizing the third time series data and the fourth time series data;
and if the correlation exists between the operation variable and the controlled variable, taking the operation variable as a target operation variable.
Wherein the second sampling interval may be the same as the first sampling interval.
In some embodiments, the verifying whether the manipulated variable and the controlled variable have a correlation using the third time series data and the fourth time series data may include:
and performing causal check on the operation variable and the controlled variable by using the first time sequence data and the second time sequence data to determine whether correlation exists.
As an alternative, a Granger causal relationship test (english: Granger causal relationship test) method may be used to perform a causal relationship check between the manipulated variable and the controlled variable, and if the causal relationship check exists between the manipulated variable and the controlled variable, it may be determined that there is a correlation between the manipulated variable and the controlled variable.
In some embodiments, the performing the causal check on the manipulated variable and the controlled variable using the first time-series data and the second time-series data may be by determining P [ cv (t)][I(t-1)]≠P[cv(t)][I-mv(t-1)]Whether the implementation is true or not is judged, wherein cv (t) represents a controlled variable of a t time step, and II (t-1) represents all information of the t-1 time step, including an operation variable and the controlled variable; i is-mv(t-1) represents all information except the manipulated variables at time step t-1.
If the inequality is established, the operating variable has no influence on the controlled variable, otherwise, the operating variable is considered to have influence on the controlled variable, and the operating variable and the controlled variable have correlation.
In addition, performing a causal check requires ensuring that the time series data is a stationary time series. Thus, in certain embodiments, the method may further comprise:
testing whether the third time series data and the fourth time series data meet stability requirements;
and carrying out differential processing on the time sequence data which does not meet the stationarity requirement until the stationarity requirement is met.
The stationarity requirement may refer to the fact that for a time series, if the mean is unchanged or the variance is unchanged, the periodicity variation is strictly eliminated.
For example, if the mean value for a time series is a constant value, the time series can be considered to be a stationary time series.
If the time series data which does not meet the stationarity requirement is subjected to differential processing. The difference processing is realized by subtracting the value of the current time step from the value of the next time step.
As another alternative, the correlation check between the manipulated variable and the controlled variable may also be implemented in other manners, for example, an a/B test (also called a split test or a bucket test) manner may be used to perform the correlation check, for Y input variables (including the manipulated variable and the controlled variable), a manipulated variable is fixed for a controlled variable, and Y input variables are tested, and if Y-1 input variables that do not include the manipulated variable have a difference in their influence, this may indicate that the manipulated variable has an influence on the controlled variable.
As can be seen from the above description, the first model may be a recurrent neural network model, since the neural network is mainly composed of an input layer, a hidden layer and an output layer, the hidden layer of the recurrent neural network is composed of network elements of chained vectors, each network element corresponds to a time step, and the hidden layer is composed of M network elements chained together for time-series data including data of M time steps. As shown in fig. 2, a network architecture diagram illustrating the deployment of hidden layers in a recurrent neural network is illustrated.
In some embodiments, when the first model is a recurrent neural network model, the training the first model using the first time-series data of each of the at least one controlled variable and the second time-series data of each of the at least one target manipulated variable may include:
and for the network unit corresponding to each time step, taking the operation data of the time step corresponding to the at least one target operation variable and the hidden state of the previous time step as the input data of the network unit, and taking the controlled data of the time step corresponding to the at least one controlled variable as the output result of the network unit, and training the recurrent neural network model.
For example, as shown in FIG. 2, assume that a target manipulated variable mv and a controlled variable cv are taken as examples, mvtIs the input of the t time step, i.e. the operating data of the target operating variable, stIs the hidden state of the t time step, obtained by the hidden state of the previous time step and the input of the current time step, wherein st＝f(Umvt+W st-1)，cvtIs the output of the t time step, i.e. the controlled data of the controlled variable, where cvt＝softmax(Vst) Where U, V and W represent model parameters, respectively, and the model parameters are shared by the network elements. The recurrent neural network can learn the nonlinear relation between the time sequence data, has memorability, parameter sharing and image-based completeness, and therefore has a good advantage in learning the nonlinear characteristics of the sequence data.
The first model obtained by training according to the data processing method described above may be used as a prediction model in model control prediction to implement automatic control, and the following describes the technical solution of the present application from the viewpoint of automatic control, as shown in fig. 3, for a flowchart of an embodiment of an automatic control method provided by the embodiment of the present application, the method may include the following steps:
301: determining at least one controlled variable and at least one target operating variable;
302: and predicting and obtaining a predicted value of the at least one controlled variable at the current time step by utilizing the first model based on the operation data of the at least one target operation variable at the previous time step.
Wherein the first model is obtained by training based on first time-series data corresponding to the at least one controlled variable acquired from historical production data and second time-series data corresponding to the at least one target operation variable.
The specific training mode of the first model may be as described above, and is not repeated herein.
303: and searching target operation data of the at least one controlled variable when the at least one controlled variable meets a preset condition based on the difference value between the predicted value and the expected value of the at least one controlled variable and the limit value of the at least one target operation variable.
304: and taking the target operation data of the at least one target operation variable as a control input value of the current time step for automatic control.
Alternatively, the target operation data of the at least one operation variable may be used as a Control input value of the current time step, and the target operation data may be automatically controlled by a Distributed Control System (DCS) in the automatic Control System. The DCS can control at least one target operation variable to operate according to the corresponding target operation data, so that the controlled data of the controlled variable are influenced, the production requirement is met, and the production efficiency is improved.
Therefore, the automatically controlling, with the target operation data of the at least one target operation variable as the control input value of the current time step, may include:
and sending the target operation data of the at least one target operation variable to a Distributed Control System (DCS), using the DCS as the control input value of the current time step, and controlling the at least one target operation variable to operate according to the control input value of the at least one target operation variable.
In the embodiment, the first model is adopted to represent the prediction model and is obtained by utilizing time series data training, so that the accuracy of the prediction model is improved, and the accuracy of the automatic control process is improved.
In some embodiments, the finding target operation data of the at least one target operation variable when the at least one controlled variable satisfies a preset condition based on a difference between the predicted value and the expected value of the at least one controlled variable and the limit value of the at least one target operation variable may include:
determining the difference value between the predicted value and the expected value of the at least one controlled variable and the limit value of the at least one target operation variable;
and searching target operation data of the at least one target operation variable when the difference value between the predicted value and the expected value of the at least one controlled variable is minimum by taking the limit value of the at least one target operation variable as a constraint condition.
Wherein, since the first model does not have a displayed mathematical expression, a PSO (Particle Swarm Optimization) algorithm or a genetic algorithm or the like may be used to find the target operation data of the at least one target operation variable. Of course, a traversal search approach may also be employed.
Therefore, in some embodiments, when training that the difference between the predicted value and the expected value of the at least one controlled variable is minimum by using the limit value of the at least one target operation variable as a constraint condition, the target operation data of the at least one target operation variable includes:
and obtaining target operation data of the at least one target operation variable when the limited value of the at least one target operation variable is used as a constraint condition and the particle swarm optimization algorithm is utilized to minimize the difference value between the predicted value and the expected value of the at least one controlled variable.
In certain embodiments, the method may further comprise:
and retraining the first model by taking the target operation data of the at least one operation variable at the current time step and the controlled data of the at least one controlled variable obtained by controlling at the current time step as historical production data.
That is, the first model can be updated all the time, so that the model accuracy and the control accuracy are ensured.
For convenience of understanding, fig. 4 shows a model prediction process schematic diagram in an automatic control process, and as can be seen from fig. 4, a first model can be trained by using historical production data of the DCS system, the first model serves as a prediction model for predicting a predicted value of a controlled variable, a desired value of the controlled variable is based on a constraint condition, a limit value of an operating variable is used as a constraint condition, a difference value between the predicted value and the desired value is minimized as a cost function, target operating data of the operating variable can be obtained through optimization, and the target operating data acts on the DCS system to realize automatic control.
The prediction and optimization process can be executed by a control device in an automatic control system, and the control device realizes automatic control of the operation variables by controlling a DCS (distributed control system).
The technical scheme of the embodiment of the application can be suitable for various industrial production scenes, and can be applied to cement production scenes, such as the production link of raw meal vertical mill of cement, and the whole production link of cement production can be automatically controlled. In a cement production business, the operating variables may include, for example, feeders, recirculation dampers, recirculation fans, cold air valves, grinding rolls, water injection valves, and the like. The controlled variables may include, for example, raw abrasive layer thickness, exit air temperature, exit water temperature, and the like. For each controlled variable, the effective operating variable, i.e. the target operating variable, can be determined according to the technical scheme of the application, and for example, the effect-proof check method of glange can be adopted to know that both the water injection valve and the circulating air door can influence the outlet mill air temperature.
From historical production data, time series data of a target operating variable and time series data of a controlled variable can be acquired and obtained, so that the time series data of the target operating variable and the time series data of the controlled variable can be utilized as a first model in advance, the first model can be used as a prediction model, the prediction model can predict and obtain a predicted value of the controlled variable based on the operating data of the target operating variable at a previous time step, and through optimization processing, when the difference value between the predicted value of the controlled variable and the expected value is minimum, the target operating data of the target operating variable can be obtained, so that automatic control of the time step can be carried out based on the target operating data of the target operating variable.
Based on the cement production business, the embodiment of the application also provides a data processing method, which comprises the following steps:
determining at least one controlled variable and at least one target operation variable related to cement production business;
acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data according to a first sampling interval;
training a first model by using the respective first time-series data of the at least one controlled variable and the respective second time-series data of the at least one target operating variable;
the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step.
The embodiment of the application also provides an automatic control method, which comprises the following steps:
determining at least one controlled variable and at least one target operation variable related to cement production business;
predicting and obtaining a predicted value of the at least one controlled variable at the current time step by utilizing a first model based on the operation data of the at least one target operation variable at the previous time step; the first model is obtained by training based on first time-series data corresponding to the at least one controlled variable and second time-series data corresponding to the at least one target operation variable acquired from historical production data;
searching target operation data of the at least one target operation variable when the at least one controlled variable meets a preset condition based on a difference value between the predicted value and the expected value of the at least one controlled variable and a limit value of the at least one target operation variable;
and taking the target operation data of the at least one target operation variable as a control input value of the current time step to automatically control the cement production.
Through the embodiment of the application, the control accuracy in the automatic cement generation process can be improved, and the energy consumption in the cement production process can be reduced.
Fig. 5 is a schematic structural diagram of an embodiment of a data processing apparatus according to an embodiment of the present application, where the apparatus may include:
a first determining module 501, configured to determine at least one controlled variable and at least one target operation variable;
a data obtaining module 502, configured to obtain, according to a first sampling interval, first time series data of the at least one controlled variable and second time series data of the at least one target operating variable from historical production data;
a model training module 503, configured to train a first model using first time-series data of the at least one controlled variable and second time-series data of the at least one target manipulated variable;
the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step.
In some embodiments, the first determining module is specifically configured to determine at least one controlled variable and at least one operating variable involved in a production service; determining at least one target manipulated variable related to the at least one controlled variable from the at least one manipulated variable.
In certain embodiments, the first determining module determining at least one target operating variable related to the at least one controlled variable from the at least one operating variable comprises:
acquiring third time series data of the controlled variables and fourth time series data of the operation variables from historical production data according to a second sampling interval for each controlled variable and each operation variable;
checking whether the correlation exists between the operation variable and the controlled variable or not by utilizing the third time series data and the fourth time series data;
and if the correlation exists between the operation variable and the controlled variable, taking the operation variable as a target operation variable.
In certain embodiments, the first determination module, using the third time series data and the fourth time series data, verifying whether there is a correlation between the manipulated variable and the controlled variable comprises:
and performing causal check on the operation variable and the controlled variable by using the first time sequence data and the second time sequence data to determine whether correlation exists.
In some embodiments, the apparatus may further comprise:
the stability testing module is used for testing whether the third time series data and the fourth time series data meet stability requirements; and carrying out differential processing on the time sequence data which does not meet the stationarity requirement until the stationarity requirement is met.
In certain embodiments, the first model is a recurrent neural network model;
the model training module is specifically configured to, for a network unit corresponding to each time step, use operation data of the time step corresponding to the at least one target operation variable and a hidden state of a previous time step as input data of the network unit, and use controlled data of the time step corresponding to the at least one controlled variable as an output result of the network unit, so as to train the recurrent neural network model.
In some embodiments, the data acquisition module is specifically configured to acquire and obtain first time series data of each of the at least one controlled variable and second time series data of each of the at least one target operating variable from historical production data within a predetermined historical time period from a current time according to a first sampling interval; wherein the first time series data comprises operational data for N time steps and the second time series data comprises controlled data for M time steps; wherein M is an integer of 1 or more.
In some embodiments, the data obtaining module may specifically acquire and obtain first time-series data corresponding to each of the at least one controlled variable and second time-series data corresponding to each of the at least one target operating variable from historical production data of the DCS at a first sampling interval.
In some embodiments, the target operation data of the at least one target operation variable at the current time step is used for sending to the DCS, the DCS takes the target operation data of the at least one target operation variable as the control input value of the current time step, and controls the at least one target operation variable to operate according to the control input value of the at least one target operation variable.
In a cement production scenario, the first determining module is specifically configured to determine at least one controlled variable and at least one target operating variable related to a cement production business.
The data processing apparatus shown in fig. 5 may execute the data processing method shown in the embodiment shown in fig. 1, and the implementation principle and the technical effect are not described again. The specific manner in which each module and unit of the data processing apparatus in the above embodiments perform operations has been described in detail in the embodiments related to the method, and will not be described in detail herein.
In one possible design, the data processing apparatus of the embodiment shown in fig. 5 may be implemented as a computing device, which may include a storage component 601 and a processing component 602 as shown in fig. 6;
the storage component 601 stores one or more computer instructions for invocation and execution by the processing component 602.
The processing component 602 is configured to:
determining at least one controlled variable and at least one target operating variable;
acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data according to a first sampling interval;
training a first model by using the respective first time-series data of the at least one controlled variable and the respective second time-series data of the at least one target operating variable;
the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step.
Among other things, the processing component 602 may include one or more processors to execute computer instructions to perform all or some of the steps of the methods described above. Of course, the processing elements may also be implemented as one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, micro-controllers, microprocessors or other electronic components configured to perform the above-described methods.
The storage component 601 is configured to store various types of data to support operations on the computing device. The memory components may be implemented by any type or combination of volatile or non-volatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disks.
Of course, a computing device may also necessarily include other components, such as input/output interfaces, communication components, and so forth.
The input/output interface provides an interface between the processing components and peripheral interface modules, which may be output devices, input devices, etc.
The communication component is configured to facilitate wired or wireless communication between the computing device and other devices, and the like.
As used herein, a "computing device" may be a remote web server, a computer networking device, a chipset, a desktop computer, a notebook computer, a workstation, or any other processing device or equipment.
The embodiment of the present application further provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a computer, the data processing method of the embodiment shown in fig. 1 may be implemented.
Fig. 7 is a schematic structural diagram of an embodiment of an automatic control device provided in an embodiment of the present application, where the automatic control device may include:
a first determining module 701, configured to determine at least one controlled variable and at least one target operation variable;
a prediction module 702, configured to predict, by using a first model, a predicted value of the at least one controlled variable at a current time step based on operation data of a previous time step of the at least one target operation variable; the first model is obtained by training based on first time-series data corresponding to the at least one controlled variable and second time-series data corresponding to the at least one target operation variable acquired from historical production data;
an optimizing module 703, configured to find target operation data of the at least one target operation variable when the at least one controlled variable meets a preset condition, based on a difference between a predicted value and an expected value of the at least one controlled variable and a limit value of the at least one target operation variable;
and the control module 704 is configured to automatically control the target operation data of the at least one target operation variable as a control input value of the current time step.
In some embodiments, the optimization module is specifically configured to determine a difference between a predicted value and an expected value of the at least one controlled variable, and a limit value of the at least one target operating variable; and searching target operation data of the at least one target operation variable when the difference value between the predicted value and the expected value of the at least one controlled variable is minimum by taking the limit value of the at least one target operation variable as a constraint condition.
In some embodiments, the optimization module uses the limit value of the at least one target operation variable as a constraint condition to find that when the difference between the predicted value and the expected value of the at least one controlled variable is minimum, the target operation data of the at least one target operation variable includes:
and obtaining target operation data of the at least one target operation variable when the limited value of the at least one target operation variable is used as a constraint condition and the particle swarm optimization algorithm is utilized to minimize the difference value between the predicted value and the expected value of the at least one controlled variable.
In some embodiments, the apparatus may further comprise:
and the training triggering module is used for retraining the first model by taking the target operation data of the at least one operation variable at the current time step and the controlled data of the at least one controlled variable obtained by controlling at the current time step as historical production data.
In some embodiments, the control module is specifically configured to send target operation data of the at least one target operation variable to a distributed control system DCS, and the DCS takes the target operation data of the at least one target operation variable as a control input value of a current time step and controls the at least one target operation variable to operate according to the control input value of the at least one target operation variable.
In a cement production scenario, the second determining module is specifically configured to determine at least one controlled variable and at least one target operation variable related to a cement production business;
the control module is specifically used for taking the target operation data of the at least one target operation variable as a control input value of the current time step to automatically control the cement production.
The data processing apparatus shown in fig. 7 may execute the automatic control method described in the embodiment shown in fig. 3, and the implementation principle and the technical effect are not described again. The specific manner in which each module and unit of the data processing apparatus in the above embodiments perform operations has been described in detail in the embodiments related to the method, and will not be described in detail herein.
In one possible design, the automatic control apparatus of the embodiment shown in fig. 7 may be implemented as a control device, which may include a storage component 801 and a processing component 802 as shown in fig. 8;
the storage component 801 stores one or more computer instructions for invocation and execution by the processing component 802.
The processing component 802 is configured to:
determining at least one controlled variable and at least one target operating variable;
predicting and obtaining a predicted value of the at least one controlled variable at the current time step by utilizing a first model based on the operation data of the at least one target operation variable at the previous time step; the first model is obtained by training based on first time-series data corresponding to the at least one controlled variable and second time-series data corresponding to the at least one target operation variable acquired from historical production data;
searching target operation data of the at least one target operation variable when the at least one controlled variable meets a preset condition based on a difference value between the predicted value and the expected value of the at least one controlled variable and a limit value of the at least one target operation variable;
and taking the target operation data of the at least one target operation variable as a control input value of the current time step for automatic control.
The processing component 802 may include one or more processors executing computer instructions to perform all or some of the steps of the methods described above. Of course, the processing elements may also be implemented as one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), controllers, micro-controllers, microprocessors or other electronic components configured to perform the above-described methods.
The storage component 801 is configured to store various types of data to support operations on the control device. The memory components may be implemented by any type or combination of volatile or non-volatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disks.
Of course, the control device may of course also comprise other components, such as input/output interfaces, communication components, etc.
The input/output interface provides an interface between the processing components and peripheral interface modules, which may be output devices, input devices, etc.
The communication component is configured to facilitate wired or wireless communication between the computing device and other devices, and the like.
An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a computer, the automatic control method of the embodiment shown in fig. 3 may be implemented.
It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.
The above-described embodiments of the apparatus are merely illustrative, and 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 modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions in the embodiments of the present application.

Claims (22)

1. A data processing method, comprising:
determining at least one controlled variable and at least one target operating variable;
acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data according to a first sampling interval;
training a first model by using the respective first time-series data of the at least one controlled variable and the respective second time-series data of the at least one target operating variable;
the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step.
2. The method according to claim 1, wherein the predicted value of the at least one target controlled variable is used for finding target operation data of the at least one target operation variable at the current time step when the at least one controlled variable meets a preset condition.
3. The method of claim 1, wherein determining at least one controlled variable and at least one target operating variable comprises:
determining at least one controlled variable and at least one operating variable involved in a production service;
determining at least one target manipulated variable related to the at least one controlled variable from the at least one manipulated variable.
4. The method of claim 3, wherein said determining at least one target manipulated variable related to said at least one controlled variable from said at least one manipulated variable comprises:
acquiring third time series data of the controlled variables and fourth time series data of the operation variables from historical production data according to a second sampling interval for each controlled variable and each operation variable;
checking whether the correlation exists between the operation variable and the controlled variable or not by utilizing the third time series data and the fourth time series data;
and if the correlation exists between the operation variable and the controlled variable, taking the operation variable as a target operation variable.
5. The method of claim 4, wherein the verifying whether the manipulated variables and the controlled variables have a correlation using the third time series data and the fourth time series data comprises:
and performing causal check on the operation variable and the controlled variable by using the first time sequence data and the second time sequence data to determine whether correlation exists.
6. The method of claim 3, further comprising:
testing whether the third time series data and the fourth time series data meet stability requirements;
and carrying out differential processing on the time sequence data which does not meet the stationarity requirement until the stationarity requirement is met.
7. The method of claim 1, wherein the first model is a recurrent neural network model;
the training a first model using the respective first time-series data of the at least one controlled variable and the respective second time-series data of the at least one target manipulated variable comprises:
and aiming at the network unit corresponding to each time step, taking the operation data of the time step corresponding to the at least one target operation variable and the hidden state of the previous time step as the input data of the network unit, and taking the controlled data of the time step corresponding to the at least one controlled variable as the output result of the network unit, and training the recurrent neural network model.
8. The method of claim 7, wherein the recurrent neural network model comprises a long-short term memory network or a gated recurrent neural network.
9. The method of claim 1, wherein the obtaining the first time-series data of the at least one controlled variable and the second time-series data of the at least one target manipulated variable from the historical production data at the first sampling interval comprises:
acquiring and obtaining first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data within a preset historical time period from the current time according to a first sampling interval; wherein the first time series data comprises operational data for N time steps and the second time series data comprises controlled data for M time steps; wherein M is an integer of 1 or more.
10. The method of claim 1, wherein the obtaining the first time-series data of the at least one controlled variable and the second time-series data of the at least one target manipulated variable from the historical production data at the first sampling interval comprises:
and acquiring first time sequence data corresponding to the at least one controlled variable and second time sequence data corresponding to the at least one target operating variable from historical production data of the Distributed Control System (DCS) according to a first sampling interval.
11. The method according to claim 2, characterized in that the target operational data of the at least one target operational variable at the current time step is used for sending to a distributed control system, DCS, which takes the target operational data of the at least one target operational variable as a control input value for the current time step and controls the operation of the at least one target operational variable in accordance with the control input value of the at least one target operational variable.
12. An automatic control method, comprising:
determining at least one controlled variable and at least one target operating variable;
predicting and obtaining a predicted value of the at least one controlled variable at the current time step by utilizing a first model based on the operation data of the at least one target operation variable at the previous time step; the first model is obtained by training based on first time-series data corresponding to the at least one controlled variable and second time-series data corresponding to the at least one target operation variable acquired from historical production data;
searching target operation data of the at least one target operation variable when the at least one controlled variable meets a preset condition based on a difference value between the predicted value and the expected value of the at least one controlled variable and a limit value of the at least one target operation variable;
and taking the target operation data of the at least one target operation variable as a control input value of the current time step for automatic control.
13. The method of claim 12, wherein the finding the target operating data of the at least one target operating variable when the at least one controlled variable satisfies a preset condition based on a difference between the predicted value and the expected value of the at least one controlled variable and the limit value of the at least one target operating variable comprises:
determining the difference value between the predicted value and the expected value of the at least one controlled variable and the limit value of the at least one target operation variable;
and searching target operation data of the at least one target operation variable when the difference value between the predicted value and the expected value of the at least one controlled variable is minimum by taking the limit value of the at least one target operation variable as a constraint condition.
14. The method according to claim 13, wherein when the limiting value of the at least one target operation variable is used as the constraint condition and the difference between the predicted value and the expected value of the at least one controlled variable is found to be minimum, the target operation data of the at least one target operation variable comprises:
and obtaining target operation data of the at least one target operation variable when the limited value of the at least one target operation variable is used as a constraint condition and the particle swarm optimization algorithm is utilized to minimize the difference value between the predicted value and the expected value of the at least one controlled variable.
15. The method of claim 12, further comprising:
and retraining the first model by taking the target operation data of the at least one operation variable at the current time step and the controlled data of the at least one controlled variable obtained by controlling at the current time step as historical production data.
16. The method of claim 12, wherein automatically controlling the target operational data of the at least one target operational variable as the control input value for the current time step comprises:
and sending the target operation data of the at least one target operation variable to a Distributed Control System (DCS), using the DCS as the control input value of the current time step, and controlling the at least one target operation variable to operate according to the control input value of the at least one target operation variable.
17. A data processing method, comprising:
determining at least one controlled variable and at least one target operation variable related to cement production business;
acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data according to a first sampling interval;
training a first model by using the respective first time-series data of the at least one controlled variable and the respective second time-series data of the at least one target operating variable;
the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step.
18. An automatic control method, comprising:
determining at least one controlled variable and at least one target operation variable related to cement production business;
predicting and obtaining a predicted value of the at least one controlled variable at the current time step by utilizing a first model based on the operation data of the at least one target operation variable at the previous time step; the first model is obtained by training based on first time-series data corresponding to the at least one controlled variable and second time-series data corresponding to the at least one target operation variable acquired from historical production data;
searching target operation data of the at least one target operation variable when the at least one controlled variable meets a preset condition based on a difference value between the predicted value and the expected value of the at least one controlled variable and a limit value of the at least one target operation variable;
and taking the target operation data of the at least one target operation variable as a control input value of the current time step to automatically control the cement production.
19. A data processing apparatus, comprising:
the first determining module is used for determining at least one controlled variable and at least one target operation variable;
the data acquisition module is used for acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data according to a first sampling interval;
the model training module is used for training a first model by utilizing first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable;
the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step.
20. An automatic control device, characterized by comprising:
the first determining module is used for determining at least one controlled variable and at least one target operation variable;
the prediction module is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step by utilizing a first model based on the operation data of the at least one target operation variable at the previous time step; the first model is obtained by training based on first time-series data corresponding to the at least one controlled variable and second time-series data corresponding to the at least one target operation variable acquired from historical production data;
the optimization module is used for searching target operation data of the at least one target operation variable when the at least one controlled variable meets a preset condition based on a difference value between a predicted value and an expected value of the at least one controlled variable and a limit value of the at least one target operation variable;
and the control module is used for taking the target operation data of the at least one target operation variable as a control input value of the current time step to carry out automatic control.
21. A computing device comprising a processing component and a storage component;
the storage component stores one or more computer instructions; the one or more computer instructions to be invoked for execution by the processing component;
the processing component is to:
determining at least one controlled variable and at least one target operating variable;
acquiring first time sequence data of the at least one controlled variable and second time sequence data of the at least one target operation variable from historical production data according to a first sampling interval;
training a first model by using the respective first time-series data of the at least one controlled variable and the respective second time-series data of the at least one target operating variable;
the first model is used for predicting and obtaining a predicted value of the at least one controlled variable at the current time step based on the operation data of the at least one target operation variable at the previous time step.
22. A control device comprising a processing component and a storage component;
the storage component stores one or more computer instructions; the one or more computer instructions to be invoked for execution by the processing component;
the processing component is to:
determining at least one controlled variable and at least one target operating variable;
predicting and obtaining a predicted value of the at least one controlled variable at the current time step by utilizing a first model based on the operation data of the at least one target operation variable at the previous time step; the first model is obtained by training based on first time-series data corresponding to the at least one controlled variable and second time-series data corresponding to the at least one target operation variable acquired from historical production data;
searching target operation data of the at least one target operation variable when the at least one controlled variable meets a preset condition based on a difference value between the predicted value and the expected value of the at least one controlled variable and a limit value of the at least one target operation variable;
and taking the target operation data of the at least one target operation variable as a control input value of the current time step for automatic control.
CN201911243442.8A 2019-12-06 2019-12-06 Data processing method, automatic control method and device Pending CN112925271A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911243442.8A CN112925271A (en) 2019-12-06 2019-12-06 Data processing method, automatic control method and device

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911243442.8A CN112925271A (en) 2019-12-06 2019-12-06 Data processing method, automatic control method and device

Publications (1)

Publication Number Publication Date
CN112925271A true CN112925271A (en) 2021-06-08

Family

ID=76161720

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911243442.8A Pending CN112925271A (en) 2019-12-06 2019-12-06 Data processing method, automatic control method and device

Country Status (1)

Country Link
CN (1) CN112925271A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114066510A (en) * 2021-11-02 2022-02-18 创优数字科技(广东)有限公司 Method and device for predicting commodity sales, computer equipment and storage medium

Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102147613A (en) * 2011-04-12 2011-08-10 黄继红 Model predictive control method and system for moisture and temperature of tobacco redryer
US20120078678A1 (en) * 2010-09-23 2012-03-29 Infosys Technologies Limited Method and system for estimation and analysis of operational parameters in workflow processes
CN107273633A (en) * 2017-06-29 2017-10-20 中南大学 Varying delay method of estimation and flow time lag method of estimation is hydrocracked between multiple operation
CN107615186A (en) * 2015-08-13 2018-01-19 华为技术有限公司 The method and apparatus of Model Predictive Control
US20180024512A1 (en) * 2016-07-25 2018-01-25 General Electric Company System modeling, control and optimization
US20180087360A1 (en) * 2016-09-26 2018-03-29 International Business Machines Corporation Controlling operation of a steam-assisted gravity drainage oil well system by adjusting controls based on forecast emulsion production
CN108921688A (en) * 2018-07-02 2018-11-30 阿里巴巴集团控股有限公司 Construct the method and device of prediction model
CN109716002A (en) * 2017-01-23 2019-05-03 彼得罗菲奥伦蒂尼有限公司 For monitoring the method for the device of adjusting gas flow and using the regulating system of the method
CN109814513A (en) * 2019-03-20 2019-05-28 杭州辛孚能源科技有限公司 A kind of catalytic cracking unit optimization method based on data model
CN109976165A (en) * 2019-04-29 2019-07-05 西安建筑科技大学 Model Predictive Control Parameter analysis of electrochemical tuning method and system towards multivariable new fractional-order system
CN110220554A (en) * 2019-06-05 2019-09-10 北京百度网讯科技有限公司 Method and apparatus for output information
US20210148213A1 (en) * 2017-08-21 2021-05-20 Landmark Graphics Corporation Neural Network Models For Real-Time Optimization of Drilling Parameters During Drilling Operations

Patent Citations (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20120078678A1 (en) * 2010-09-23 2012-03-29 Infosys Technologies Limited Method and system for estimation and analysis of operational parameters in workflow processes
CN102147613A (en) * 2011-04-12 2011-08-10 黄继红 Model predictive control method and system for moisture and temperature of tobacco redryer
CN107615186A (en) * 2015-08-13 2018-01-19 华为技术有限公司 The method and apparatus of Model Predictive Control
US20180024512A1 (en) * 2016-07-25 2018-01-25 General Electric Company System modeling, control and optimization
US20180087360A1 (en) * 2016-09-26 2018-03-29 International Business Machines Corporation Controlling operation of a steam-assisted gravity drainage oil well system by adjusting controls based on forecast emulsion production
CN109716002A (en) * 2017-01-23 2019-05-03 彼得罗菲奥伦蒂尼有限公司 For monitoring the method for the device of adjusting gas flow and using the regulating system of the method
CN107273633A (en) * 2017-06-29 2017-10-20 中南大学 Varying delay method of estimation and flow time lag method of estimation is hydrocracked between multiple operation
US20210148213A1 (en) * 2017-08-21 2021-05-20 Landmark Graphics Corporation Neural Network Models For Real-Time Optimization of Drilling Parameters During Drilling Operations
CN108921688A (en) * 2018-07-02 2018-11-30 阿里巴巴集团控股有限公司 Construct the method and device of prediction model
CN109814513A (en) * 2019-03-20 2019-05-28 杭州辛孚能源科技有限公司 A kind of catalytic cracking unit optimization method based on data model
CN109976165A (en) * 2019-04-29 2019-07-05 西安建筑科技大学 Model Predictive Control Parameter analysis of electrochemical tuning method and system towards multivariable new fractional-order system
CN110220554A (en) * 2019-06-05 2019-09-10 北京百度网讯科技有限公司 Method and apparatus for output information

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
李勇;吴敏;曹卫华;赖旭芝;王春生;: "烧结矿质量级联集成智能预测模型", 仪器仪表学报, no. 08 *
王华秋;廖晓峰;: "非线性量化小脑模型神经网络在溶出循环母液配比模型中的应用", 信息与控制, no. 01, 15 February 2010 (2010-02-15) *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114066510A (en) * 2021-11-02 2022-02-18 创优数字科技(广东)有限公司 Method and device for predicting commodity sales, computer equipment and storage medium

Similar Documents

Publication Publication Date Title
US20200348632A1 (en) Heuristic method of automated and learning control, and building automation systems thereof
Gao et al. Cloud-enabled prognosis for manufacturing
Jovanović et al. Ensemble of various neural networks for prediction of heating energy consumption
Tan et al. A three-step machine learning framework for energy profiling, activity state prediction and production estimation in smart process manufacturing
CN111210024A (en) Model training method and device, computer equipment and storage medium
Khosravi et al. An optimized mean variance estimation method for uncertainty quantification of wind power forecasts
US20190392363A1 (en) Predicting arrival time of components based on historical receipt data
KR102103006B1 (en) Method and Apparatus for Operating Optimal of Equipment based on Machine Learning Model
CN113609779B (en) Modeling method, device and equipment for distributed machine learning
CN111033552B (en) Cognitive energy assessment of non-invasive sensors in a thermal energy fluid transfer system
Chen et al. Optimal variability sensitive condition-based maintenance with a Cox PH model
CN116089870A (en) Industrial equipment fault prediction method and device based on meta-learning under small sample condition
Hammami et al. Neural networks for online learning of non-stationary data streams: a review and application for smart grids flexibility improvement
CN115660226B (en) Power load prediction model construction method and digital twin-based construction device
Jahani et al. Stochastic prognostics under multiple time-varying environmental factors
Lutska et al. Forecasting the efficiency of the control system of the technological object on the basis of neural networks
Zhang et al. Aeroengines remaining useful life prediction based on improved C-loss ELM
CN112925271A (en) Data processing method, automatic control method and device
Bureš et al. Forming ensembles at runtime: a machine learning approach
US20210350282A1 (en) Managing an installed base of artificial intelligence modules
Acebes et al. Statistical learning techniques for project control
He-Hua et al. A multiple criteria decision analysis method for alternative assessment results obeying a particular distribution and application
Chen Reconstructing genetic regulatory networks using two-step algorithms with the differential equation models of neural networks
Rojek et al. Industrial control system based on data processing
Musa A Hybrid Modern and Classical Algorithm for Indonesian Electricity Demand Forecasting

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