CN111539545B - Oil temperature prediction method and device and computer readable storage medium - Google Patents
Oil temperature prediction method and device and computer readable storage medium Download PDFInfo
- Publication number
- CN111539545B CN111539545B CN201910057412.1A CN201910057412A CN111539545B CN 111539545 B CN111539545 B CN 111539545B CN 201910057412 A CN201910057412 A CN 201910057412A CN 111539545 B CN111539545 B CN 111539545B
- Authority
- CN
- China
- Prior art keywords
- oil temperature
- historical
- predicted
- data
- oil
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000003860 storage Methods 0.000 title claims abstract description 25
- 238000012549 training Methods 0.000 claims abstract description 70
- 238000013528 artificial neural network Methods 0.000 claims abstract description 66
- 238000011144 upstream manufacturing Methods 0.000 claims abstract description 62
- 238000012545 processing Methods 0.000 claims description 24
- 230000002159 abnormal effect Effects 0.000 claims description 15
- 238000001914 filtration Methods 0.000 claims description 14
- 238000004422 calculation algorithm Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 8
- 238000003062 neural network model Methods 0.000 claims description 6
- 230000003111 delayed effect Effects 0.000 claims description 5
- 238000012886 linear function Methods 0.000 claims description 4
- 230000009467 reduction Effects 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 2
- 239000003208 petroleum Substances 0.000 abstract description 2
- 239000003921 oil Substances 0.000 description 281
- 238000012360 testing method Methods 0.000 description 22
- 238000010586 diagram Methods 0.000 description 11
- 230000008569 process Effects 0.000 description 11
- 238000012795 verification Methods 0.000 description 8
- 230000001932 seasonal effect Effects 0.000 description 7
- 238000013507 mapping Methods 0.000 description 4
- 230000008021 deposition Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000004927 fusion Effects 0.000 description 3
- 238000004519 manufacturing process Methods 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 239000010779 crude oil Substances 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000010438 heat treatment Methods 0.000 description 2
- 238000011022 operating instruction Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000001556 precipitation Methods 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- 241000287196 Asthenes Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004581 coalescence Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 238000009833 condensation Methods 0.000 description 1
- 230000005494 condensation Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 239000013078 crystal Substances 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000000704 physical effect Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 239000002689 soil Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000002945 steepest descent method Methods 0.000 description 1
- 230000036962 time dependent Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06N3/084—Backpropagation, e.g. using gradient descent
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Economics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Tourism & Hospitality (AREA)
- Marketing (AREA)
- General Business, Economics & Management (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Life Sciences & Earth Sciences (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Development Economics (AREA)
- Primary Health Care (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Public Health (AREA)
- Quality & Reliability (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Operations Research (AREA)
- Water Supply & Treatment (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses an oil temperature prediction method and device and a computer readable storage medium, and belongs to the field of petroleum transportation. The method comprises the following steps: acquiring a training data set, wherein the training data set comprises a plurality of data samples; calling an oil temperature prediction model based on a plurality of data samples to perform oil temperature prediction to obtain a plurality of predicted historical oil temperatures, wherein the oil temperature prediction model is obtained by training a back propagation neural network by adopting a training data set; subtracting the plurality of predicted historical oil temperatures from the historical downstream incoming oil temperatures of the plurality of predicted historical oil temperatures respectively to obtain a residual value sequence; calling an autoregressive moving average model to predict a residual value of the residual value sequence at the next moment; calling an oil temperature prediction model to perform oil temperature prediction based on oil temperature data to be predicted to obtain predicted oil temperature at the next moment, wherein the oil temperature data to be predicted comprises upstream oil temperature, upstream ground temperature, downstream station-entering ground temperature and output; and adding the residual value at the next moment and the predicted oil temperature at the next moment to obtain a predicted value of the inbound oil temperature at the next moment.
Description
Technical Field
The invention relates to the field of petroleum transportation, in particular to a method and a device for predicting oil temperature and a computer readable storage medium.
Background
When the oil temperature along the line is lower than the wax precipitation point of the transported oil product in the operation process of the hot oil pipeline, the radial temperature difference of the oil wall and the generated wax crystal coalescence effect enable the wax precipitation on the pipe wall to form a wax deposition layer. The wax deposition layer not only can cause the effective sectional area of the pipeline to be reduced, but also can cause the friction along the pipeline to be increased, the power cost to be increased and the production unit consumption to be increased; meanwhile, the increase of the wax deposition layer also affects the operation safety of the pipeline, and can cause wax to block a condensation pipe in serious cases. However, if the oil temperature along the pipeline is too high in the operation process, the economical efficiency of the pipeline operation is affected. Therefore, the prediction of the inbound oil temperature becomes the primary task for realizing the safe and optimal operation of the hot oil pipeline.
At present, the prediction of the inbound oil temperature mainly adopts the following modes: and (3) predicting the steady-state oil temperature of the hot oil pipeline by utilizing a Suhoff oil temperature calculation formula in combination with a total heat transfer coefficient (K value) based on the upstream oil temperature, the upstream ground temperature, the downstream station entering ground temperature and the output.
However, when the hot oil pipeline is actually operated, the ground temperature along the pipeline changes due to seasonal changes, the temperature difference between the hot oil pipeline and the soil changes greatly, and particularly the K value changes greatly due to dynamic oil temperature changes during process adjustment, so that the oil temperature predicted by adopting a Suhoff oil temperature calculation formula cannot accurately guide the actual operation of the pipeline.
Disclosure of Invention
The embodiment of the invention provides an oil temperature prediction method and device and a computer readable storage medium, which can realize accurate prediction of oil temperature. The technical scheme is as follows:
in one aspect, an embodiment of the present invention provides a method for predicting oil temperature, where the method includes:
acquiring a training data set, wherein the training data set comprises a plurality of data samples, and each data sample comprises time, historical upstream oil temperature, historical upstream ground temperature, historical downstream station entering ground temperature, historical output and historical downstream station entering oil temperature;
calling an oil temperature prediction model to predict oil temperature based on historical upstream oil temperature, historical upstream ground temperature, historical downstream station entering ground temperature and historical output in a plurality of data samples to obtain a plurality of predicted historical oil temperatures, wherein the plurality of data samples are continuously extracted within a period of time, and the oil temperature prediction model is obtained by training a back propagation neural network by adopting the training data set;
subtracting the plurality of predicted historical oil temperatures from historical downstream incoming oil temperatures of the plurality of predicted historical oil temperatures respectively to obtain a residual value sequence;
calling an autoregressive moving average model to predict a residual value of the residual value sequence at the next moment;
calling the oil temperature prediction model based on oil temperature data to be predicted to perform oil temperature prediction to obtain predicted oil temperature at the next moment, wherein the oil temperature data to be predicted comprises upstream oil temperature, upstream ground temperature, downstream station-entering ground temperature and output;
and adding the residual value at the next moment and the predicted oil temperature at the next moment to obtain a predicted value of the inbound oil temperature at the next moment.
In an implementation manner of the embodiment of the present invention, the acquiring a training data set includes:
extracting historical oil temperature data at preset time intervals;
filtering the extracted historical oil temperature data to remove abnormal data;
and converting the filtered historical oil temperature data into a preset format to obtain a plurality of data samples.
In an implementation manner of the embodiment of the present invention, before the calling the autoregressive moving average model to predict the residual value of the residual value sequence at the next time, the method further includes:
and carrying out noise reduction processing on the residual value sequence.
In an implementation manner of the embodiment of the present invention, the method further includes:
and taking the predicted value of the inbound oil temperature at the next moment as the predicted value of the inbound oil temperature delayed for a period of time at the next moment.
In another aspect, an embodiment of the present invention provides an oil temperature prediction apparatus, where the apparatus includes:
the system comprises an acquisition module, a data acquisition module and a data processing module, wherein the acquisition module is used for acquiring a training data set, the training data set comprises a plurality of data samples, and each data sample comprises time, historical upstream oil temperature, historical upstream ground temperature, historical downstream arrival ground temperature, historical output and historical downstream arrival oil temperature;
the first prediction module is used for calling an oil temperature prediction model to predict oil temperature based on historical upstream oil temperature, historical upstream geothermal temperature, historical downstream station-entering geothermal temperature and historical output quantity in a plurality of data samples to obtain a plurality of predicted historical oil temperatures, the plurality of data samples are continuously extracted within a period of time, and the oil temperature prediction model is obtained by training a back propagation neural network by adopting the training data set;
the calculation module is used for subtracting the plurality of predicted historical oil temperatures from the historical downstream incoming oil temperatures of the plurality of predicted historical oil temperatures respectively to obtain a residual value sequence;
the second prediction module is used for calling an autoregressive moving average model to predict a residual value of the residual value sequence at the next moment;
the first prediction module is further used for calling the oil temperature prediction model to perform oil temperature prediction based on oil temperature data to be predicted to obtain predicted oil temperature at the next moment, and the oil temperature data to be predicted comprises upstream oil temperature, upstream ground temperature, downstream station-entering ground temperature and output;
the calculation module is further configured to add the residual value at the next time and the predicted oil temperature at the next time to obtain a predicted value of the inbound oil temperature at the next time.
In an implementation manner of the embodiment of the present invention, the obtaining module is configured to extract historical oil temperature data at predetermined time intervals; filtering the extracted historical oil temperature data, and removing abnormal data; and converting the filtered historical oil temperature data into a preset format to obtain a plurality of data samples.
In an implementation manner of the embodiment of the present invention, the second prediction module is further configured to perform noise reduction processing on the residual value sequence before the autoregressive moving average model is called to predict the residual value at the next time of the residual value sequence.
In an implementation manner of the embodiment of the present invention, the apparatus further includes: and the processing module is used for taking the predicted value of the incoming oil temperature at the next moment as the predicted value of the incoming oil temperature delayed for a period of time at the next moment.
In another aspect, an embodiment of the present invention provides an oil temperature prediction apparatus, where the apparatus includes: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute executable instructions stored in the memory to implement any of the foregoing oil temperature prediction methods.
In another aspect, an embodiment of the present invention provides a computer-readable storage medium, wherein instructions, when executed by a processor of an oil temperature prediction device, enable the oil temperature prediction device to perform any one of the oil temperature prediction methods described above.
The technical scheme provided by the embodiment of the invention has the following beneficial effects:
firstly, a training data set is obtained, the training data set comprises a plurality of data samples, each data sample comprises time, historical upstream oil temperature, historical upstream ground temperature, historical downstream arrival ground temperature, historical output and historical downstream arrival oil temperature, and an arrival oil temperature prediction model is used for predicting the arrival oil temperature according to the training samples in the training data set to obtain a plurality of predicted historical oil temperatures; then, a residual value sequence is obtained by predicting the historical oil temperature and the historical downstream inbound oil temperature; predicting a residual value of the residual value sequence at the next moment by using an autoregressive moving average model; and then, predicting the predicted oil temperature at the next time by using an oil temperature prediction model, and adding the residual value at the next time and the predicted oil temperature at the next time to obtain a predicted value of the incoming oil temperature at the next time. The oil temperature is predicted in a mode of combining a neural network model and an autoregressive moving average model, and the residual value caused by seasonal variation characteristics of the oil temperature can be calculated through the autoregressive moving average model in the scheme, so that the predicted value of the neural network is compensated, namely, the influence of seasonal factors is considered in the prediction process, the accuracy of oil temperature prediction is ensured, and the prediction accuracy can be controlled to be +/-5 ℃.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the description below are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of a method for predicting an oil temperature according to an embodiment of the present invention;
FIG. 2 is a flow chart of another method for predicting oil temperature provided by an embodiment of the present invention;
FIG. 3 is a topological structure diagram of a BP neural network;
FIG. 4 is a schematic diagram illustrating predicted results of oil temperatures provided by an embodiment of the present invention;
FIG. 5 is a schematic representation of residual data for the oil temperature prediction of FIG. 4;
FIG. 6 is a schematic diagram illustrating another predicted oil temperature according to an embodiment of the present invention;
FIG. 7 is a schematic representation of residual data for a prediction of oil temperature after a delay adjustment;
FIG. 8 is a graph of predicting a pipeline oil temperature trend using only a BP neural network;
FIG. 9 is a graph of error trend of predicting oil temperature of a pipeline by using a BP neural network only;
FIG. 10 is a graph of a BP neural network and an ARMA model predicting pipeline oil temperature trends when test data sets are used;
FIG. 11 is a BP neural network and ARMA model predicted pipeline oil temperature error histogram when a test data set is used;
FIG. 12 is a graph of a BP neural network and an ARMA model predicting pipeline oil temperature trends when training data sets are used;
FIG. 13 is a BP neural network and ARMA model predicted pipeline oil temperature error histogram when a training data set is used;
FIG. 14 is a graph of a BP neural network and an ARMA model predicting a pipeline oil temperature error trend when a training data set is used;
FIG. 15 is a graph of BP neural network and ARMA model predicted pipeline oil temperature trends when test data sets were used;
FIG. 16 is a graph of predicted versus true value trends for # 2 hot station training data set;
FIG. 17 is a comparison graph of the 2# hot station training error;
FIG. 18 is a graph of predicted versus true value trends for # 2 hot station test data set;
FIG. 19 is a comparison graph of 2# hot station test error;
FIG. 20 is a graph of predicted versus true value trends for the 4# end station training data set;
FIG. 21 is a comparison graph of 4# end station training errors;
FIG. 22 is a graph of predicted versus true value trends for the 4# end station test data set;
FIG. 23 is a comparison graph of 4# end station error;
FIG. 24 is a schematic structural diagram of an oil temperature prediction apparatus according to an embodiment of the present invention;
fig. 25 is a block diagram of a configuration of an oil temperature prediction device according to an embodiment of the present disclosure.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention will be described in detail with reference to the accompanying drawings.
Fig. 1 is a flowchart of a method for predicting oil temperature according to an embodiment of the present invention. Referring to fig. 1, the method includes:
step 101: a training data set is obtained.
In this step, the training data set includes a plurality of data samples, each data sample including time, historical upstream oil temperature, historical upstream ground temperature, historical downstream inbound ground temperature, historical throughput, and historical downstream inbound oil temperature.
In the embodiment of the invention, the training data set is obtained by preprocessing historical oil temperature data, and the historical oil temperature data is obtained from three running hot oil pipelines. After historical oil temperature data of the three hot oil pipelines are obtained, the historical oil temperature data are preprocessed to obtain a data set. And (3) performing model training by using 70% of data samples in the data set as a training data set, and performing model verification by using 30% of data samples as a test data set.
Wherein the preprocessing of the historical oil temperature data may include: the filtering process and the formatting process. The filtering processing refers to filtering abnormal values in the data, and the format processing refers to converting the historical oil temperature data into a preset format.
Step 102: and calling an oil temperature prediction model to predict the oil temperature based on the historical upstream oil temperature, the historical upstream ground temperature, the historical downstream station entering ground temperature and the historical output in the plurality of data samples, so as to obtain a plurality of predicted historical oil temperatures.
Wherein the plurality of data samples are continuously extracted over a period of time.
The oil temperature prediction model is obtained by training a Back Propagation (BP) neural network by adopting a training data set, and the BP neural network is a multilayer feedforward network trained according to an error Back Propagation algorithm. The BP neural network is able to learn and store a large number of input-output pattern mappings without prior disclosure of mathematical equations describing such mappings. The learning rule is to use the steepest descent method to continuously adjust the weight and the threshold value of the network through back propagation so as to minimize the error square sum of the network.
Historical upstream oil temperature, historical upstream ground temperature, historical downstream station-entering ground temperature and historical output in the data samples are adopted, and the historical oil temperature can be predicted in an oil temperature prediction model.
Step 103: and subtracting the plurality of predicted historical oil temperatures from the historical downstream incoming oil temperatures of the plurality of predicted historical oil temperatures respectively to obtain a residual value sequence.
The predicted historical oil temperatures in step 102 each correspond to a real historical oil temperature, i.e., a historical downstream inbound oil temperature. And subtracting each predicted historical oil temperature from the corresponding historical downstream incoming station oil temperature to obtain the error between the oil temperature predicted by the oil temperature prediction model and the true value, namely a residual value.
And predicting the predicted historical oil temperatures at different times according to the time sequence, namely, subtracting the predicted historical oil temperatures from the true values to obtain residual values at different times, and arranging the residual values according to time to obtain a residual value sequence.
Step 104: and calling an autoregressive moving average model to predict the residual value of the residual value sequence at the next moment.
Among them, Auto-Regressive Moving Average Model (ARMA) is an important method for researching time series, and plays an important role in researching data with seasonal variation characteristics.
Therefore, the time-dependent residual value sequence is input into the autoregressive moving average model, and the residual value at the next moment can be predicted so as to complete the prediction of the downstream inbound oil temperature at the next moment.
Step 105: and calling an oil temperature prediction model to perform oil temperature prediction based on the oil temperature data to be predicted to obtain predicted oil temperature at the next moment, wherein the oil temperature data to be predicted comprises upstream oil temperature, upstream ground temperature, downstream station-entering ground temperature and output.
The oil temperature at the next moment can be predicted in the oil temperature prediction model by adopting the upstream oil temperature, the upstream ground temperature, the downstream station-entering ground temperature and the output in the oil temperature data to be predicted.
Step 106: and adding the residual value at the next moment and the predicted oil temperature at the next moment to obtain a predicted value of the inbound oil temperature at the next moment.
And correcting the predicted oil temperature of the back propagation neural network at the next moment by using the residual value of the back propagation neural network at the next moment calculated by the autoregressive moving average model to obtain the predicted value of the inbound oil temperature at the next moment.
In the embodiment of the invention, a training data set is obtained firstly, the training data set comprises a plurality of data samples, each data sample comprises time, historical upstream oil temperature, historical upstream ground temperature, historical downstream arrival ground temperature, historical output and historical downstream arrival oil temperature, and an arrival oil temperature prediction is carried out by using an oil temperature prediction model according to the training samples in the training data set to obtain a plurality of predicted historical oil temperatures; then, a residual value sequence is obtained by predicting the historical oil temperature and the historical downstream inbound oil temperature; predicting a residual value of the residual value sequence at the next moment by using an autoregressive moving average model; and then, predicting the predicted oil temperature at the next time by using an oil temperature prediction model, and adding the residual value at the next time and the predicted oil temperature at the next time to obtain the predicted value of the incoming oil temperature at the next time. The oil temperature is predicted in a mode of combining a neural network model and an autoregressive moving average model, and the residual value caused by seasonal variation characteristics of the oil temperature can be calculated through the autoregressive moving average model in the scheme, so that the predicted value of the neural network is compensated, namely, the influence of seasonal factors is considered in the prediction process, the accuracy of oil temperature prediction is ensured, and the prediction accuracy can be controlled to be +/-5 ℃.
FIG. 2 is a flow chart of another oil temperature prediction method according to an embodiment of the present invention. Referring to fig. 2, the method includes:
step 201: historical oil temperature data is extracted at predetermined time intervals.
In this step, the historical oil temperature data is the historical oil temperature data acquired from the three operating hot oil pipes. For example, data of each pipe section of three pipelines from 3 months in 2015 to 6 months in 2017 are selected, historical oil temperature data are extracted from the data at intervals of half an hour, and the historical oil temperature data of each pipeline are about 4 ten thousand. Wherein each historical oil temperature data comprises time, historical upstream oil temperature, historical upstream ground temperature, historical downstream station entering ground temperature, historical output and historical downstream station entering oil temperature.
Step 202: and filtering the extracted historical oil temperature data, and removing abnormal data.
Wherein, the filtering process can be realized by two modes of manual processing and automatic machine processing:
manual treatment: firstly, the historical oil temperature data is displayed, the operating instruction of a worker is received, and the abnormal data is labeled according to the operating instruction. Namely, through visual preview of the historical oil temperature data, the worker finds out the abnormal data and marks the abnormal data. The abnormal data is data exceeding a normal value interval or empty data which is not collected. Then, the marked data is cleaned, for example, the abnormal data is replaced by a set value, or the empty data is filled, or the abnormal data is deleted. By the steps, invalid deviation values and non-collected null values are reasonably replaced and filled, and the quality and the effectiveness of data are guaranteed.
Automatic processing by a machine: after the historical oil temperature data are obtained through extraction, the common deviation values with excessive temperature in the time periods before and after the deviation are filtered out in a filtering mode.
Step 203: and converting the filtered historical oil temperature data into a preset format to obtain a plurality of data samples.
After the filtering processing, the historical oil temperature data are stored in a uniform preset format, and each piece of historical oil temperature data becomes a data sample so as to facilitate the use of the model. The format conversion here may be a normalization process, for example, normalizing the temperature data to data between 0 and 1 to obtain data samples.
Each data sample includes time, historical upstream oil temperature, historical upstream ground temperature, historical downstream inbound ground temperature, historical throughput, and historical downstream inbound oil temperature.
The plurality of data samples constitute a data set, and model training is performed using 70% of the data samples in the data set as a training data set, and model verification is performed using 30% of the data samples in the data set as a test data set.
Step 204: and training a back propagation neural network by adopting a data set to obtain an oil temperature prediction model.
Before step 204, a model of the back propagation neural network needs to be constructed, and then the data set is input to the back propagation neural network for training.
When a model of a back propagation neural network is built, preliminary analysis needs to be carried out on historical oil temperature data, the degree of correlation of characteristic data to an overall regression algorithm is judged, then the neural network model is built, parameters of the built algorithm are adjusted and optimized by using a neural network prediction result, and therefore the optimization of overall prediction capability is achieved.
Among other things, BP neural networks can learn and store a large number of input-output pattern mappings without the need to pre-develop mathematical equations describing such mappings. Fig. 3 is a topology structure diagram of a BP neural network, and referring to fig. 3, the BP neural network includes an input layer (input layer), a hidden layer (hide layer), and an output layer (output layer). The BP neural network is trained through a BP algorithm, the basic idea of the BP algorithm is a least square method, and the weight and the threshold of the network are adjusted according to a rapid gradient descent method, so that the error square sum of the network is minimum.
In the embodiment of the invention, the BP neural network can adopt a rapid learning algorithm (RLS) based on Kalman filtering, so that the learning speed of the forward neural network is obviously improved, the learning parameters have certain robustness, and the probability of the network falling into local minimum is lower due to the self-adaptive characteristic of the Kalman filtering.
Step 205: and calling an oil temperature prediction model to predict the oil temperature based on the historical upstream oil temperature, the historical upstream ground temperature, the historical downstream station entering ground temperature and the historical output in the plurality of data samples, so as to obtain a plurality of predicted historical oil temperatures.
Where the plurality of data samples are taken continuously over a period of time.
In step 204, an oil temperature prediction model is obtained through training, and the historical oil temperature can be predicted in the oil temperature prediction model by using the historical upstream oil temperature, the historical upstream geothermal temperature, the historical downstream arrival geothermal temperature and the historical output in the data sample.
Step 206: and subtracting the plurality of predicted historical oil temperatures from the historical downstream incoming oil temperatures of the plurality of predicted historical oil temperatures respectively to obtain a residual value sequence.
And predicting the historical oil temperature of the inbound station by adopting the back propagation neural network obtained by training in the step 204, and subtracting the predicted historical oil temperature from the historical oil temperature of the inbound station to obtain a residual value after obtaining the predicted historical oil temperature. And predicting the predicted historical oil temperatures at different times according to the time sequence, namely, subtracting the predicted historical oil temperatures from the true values to obtain residual values at different times, and arranging the residual values according to time to obtain a residual value sequence. Since a plurality of data samples are successively extracted over a period of time, the sequence of residual values here also corresponds to successive extraction time points over a period of time.
Step 207: and calling the autoregressive moving average model to predict the residual value of the residual value sequence at the next moment.
In an embodiment of the invention, the data predicted for the neural network is further processed through a time series. The method for processing dynamic random data based on the linear model established by the time sequence mainly comprises an autoregressive model (AR), a Moving Average (MA) model, an autoregressive moving average (ARMA) model and the like.
Among them, the ARMA model is an important method for studying time series, and plays an important role in studying data having seasonal variation characteristics. The predicted value of the ARMA model is a linear function of the historical value and the random error, and the value of the ARMA model at the time t is y t It is not only related to its own value at its previous time, but is also affected by disturbances entering the system at the previous time, i.e. an autoregressive moving average system, whose mathematical model is ARMA (n, m):
in the formula: real parameter
Is an autoregressive coefficient; real parameter theta j (j is more than or equal to 1 and less than or equal to m) is a moving average coefficient; { a t Is a white noise sequence. Formula (1) is called an n-order autoregressive m-order moving average model, a post-shift operator B is introduced, and formula (1) can be expressed as:
in the formula:
θ(B)＝1-θ 1 B-…-θ n B q suppose that
And θ (B) is a formless equation. When p is 0, ARMA (p, q), i.e., ma (q), model; when q is 0, the ARMA (p, q) model is the ar (p) model. The ar (p) model and the ma (q) model are special cases of the ARMA (p, q) model, both of which are referred to as ARMA (p, q) model. The parameters in the formulas (1) and (2) can be automatically assigned through a model after inputting the residual value sequence under the condition of setting a threshold range. Based on ARMA time sequence characteristics, the nonlinearity of the BP neural network in the oil temperature prediction process is combined, and the method is favorableAnd performing time sequence linear regression on the BP neural network prediction error by using an ARMA model, and improving the accuracy of a prediction result after model fusion.
As the pipeline oil temperature prediction has a strong nonlinear relation, and the BP neural network prediction result has the characteristics of poor stability, nonlinear prediction and linear prediction of an ARMA model, the ARMA and BP neural network fusion model is established to realize the oil temperature prediction. The model is considered as a linear autocorrelation and nonlinear component, and is as follows:
y t ＝N t +L t (3)
in the formula, N t Is a non-linear part, L t Is a linear portion. The non-linear part is obtained by a BP neural network, the linear part is a residual error between a true value and a predicted value and is obtained by an ARMA model, and the residual error at the moment t is as follows:
in the formula (I), the compound is shown in the specification,
the predicted value of the BP neural network is used for time t. The prediction accuracy of the BP neural network is improved by using ARMA simulation residual errors, namely linear relation correction. For n input nodes, the residual ARMA neural network model is:
e t ＝f(e t-1 ，e t-2 ，…，e t-n )+ε t (5)
in the formula: f is a linear function, ε, determined by the ARMA model t For random error (i.e. error e at time t predicted by BP neural network) t Difference predicted by ARMA model), f (e) in equation (5) t-1 ，e t-2 ，…，e t-n ) Is a practical application of equation (2). Record residual prediction as
Then equation (3) is:
it can be seen that the model fusion model described above includes two steps. A first step of analyzing the predicted non-linear part using a BP neural network (steps 205, 208); in the second step, the residual error generated by the predicted BP neural network is simulated using the ARMA model (steps 206, 207).
Step 208: and calling an oil temperature prediction model to perform oil temperature prediction based on the oil temperature data to be predicted to obtain predicted oil temperature at the next moment, wherein the oil temperature data to be predicted comprises upstream oil temperature, upstream geothermal temperature, downstream station-entering geothermal temperature and output.
In step 204, an oil temperature prediction model is obtained through training, and the oil temperature at the next time can be predicted in the oil temperature prediction model by using the upstream oil temperature, the upstream ground temperature, the downstream station-entering ground temperature and the output in the oil temperature data to be predicted.
Step 209: and adding the residual value at the next moment and the predicted oil temperature at the next moment to obtain a predicted value of the inbound oil temperature at the next moment.
In the embodiment of the invention, the ARMA model is used for predicting the residual error generated by BP neural network prediction, thereby reducing error jump and improving the accuracy and generalization of model prediction.
The oil temperature prediction method provided by the embodiment of the invention is explained by experiments as follows:
and acquiring data in the selected time period from the database, reading a group of data every half hour to form a data sample, and predicting the inbound oil temperature. Fig. 4 is a schematic diagram of a predicted oil temperature result provided by an embodiment of the present invention, and referring to fig. 4, wherein a dotted line is a predicted value, and a solid line is a real value. It can be seen that the predicted values can be better fitted in the detailed part, except for the presence of salient values in part of the data. The reason for the occurrence of the saliency values in fig. 4 is random noise generated in time series, resulting in abnormal data in the residual values.
Fig. 5 is a schematic diagram of residual data for the oil temperature prediction of fig. 4. Referring to fig. 5, it can be seen that there is a small portion of clearly abnormal data (caused by random noise) and most of the data results are within ± 0.5 ℃.
Therefore, before calling the autoregressive moving average model to predict the residual value at the next time of the residual value sequence, the method may further include: and carrying out noise reduction processing on the residual value sequence. The noise-reduced result is more accurate, and more residual errors enter the range of [ -0.5, 0.5 ].
Fig. 6 is a schematic diagram of another oil temperature prediction result provided by an embodiment of the present invention, referring to fig. 6, data in a selected time period is obtained from a database, a group of data is read every day to form a data sample, and an inbound oil temperature prediction is performed, where the result is shown in fig. 6, where a dotted line is a predicted value and a solid line is a true value.
According to the lines of the two square regions in fig. 6, the predicted value is entirely advanced compared with the true value, and the delay phenomenon is more obvious when the oil temperature changes. The analysis shows that the upstream furnace start-stop operation leads to the reconstruction of a pipe section temperature field, and the downstream station-entering oil temperature needs a period of time to be constant.
Accordingly, the method may further comprise: and taking the predicted inbound oil temperature value at the next time as the predicted inbound oil temperature value after delaying for a period of time at the next time. Namely, the predicted inbound oil temperature value at the next moment is subjected to delay processing and is used as the predicted inbound oil temperature value after a period of time delay. Wherein, the period of time can be 2-3 days, and the prediction deviation is obviously reduced after the delay adjustment, and the result is shown in fig. 7. The instantaneous data of the furnace shutdown/start-up working condition of the corresponding pipe section with the grasped data are used for verification, so that the temperature change of the downstream in about 14 hours can be obtained when the furnace shutdown/start-up working condition exists at the upstream; and after about 2-3 days, the stability is achieved, so the delay adjustment is delayed for 2-3 days.
The effect of the oil temperature prediction method provided by the embodiment of the invention is verified by combining the following examples:
in order to verify the accuracy of the oil temperature prediction model of the pipeline, the actual production data of the HY hot oil pipeline is selected to perform model verification. An HY hot oil pipeline is provided with a 1# first station, a 2# hot station, a 3# hot station and a 4# last station 4 station yard, and high wax-containing crude oil is mainly output. According to the ground temperature and the physical properties of the oil along the line (see table 1), the pipeline adopts three different processes of comprehensive heat treatment, heat treatment and normal-temperature conveying.
TABLE 1
Relevant parameters from 3# hot station to 4# last station 2015 in 3, 4 and 5, 7 days of the HY crude oil pipeline are selected as model verification data. And preprocessing the data to obtain a data set, and verifying a pipeline oil temperature prediction model by using the data set.
The pipeline oil temperature is predicted by adopting a BP neural network firstly:
fig. 8 is a graph of predicting the oil temperature trend of the pipeline only by using the BP neural network, in fig. 8, a dotted line is a predicted value, a solid line is a true value, a part in front of a vertical line is a result of prediction by using a training data set, and a part behind the vertical line is a result of prediction by using a test data set. FIG. 9 is a graph of the oil temperature error trend predicted by using only the BP neural network, where the part in front of the vertical line is the error predicted by using the training data set, and the part behind the vertical line is the error predicted by using the test data set. Referring to fig. 8 and 9, it can be seen that the coincidence degree of the real value and the predicted value of the training data set is high, but the deviation between the real value and the predicted value of the testing data set is large and reaches nearly 10 ℃ at most, so that the pipeline oil temperature prediction model established by only adopting the BP neural network is verified, and the prediction accuracy is low.
BP neural network and ARMA model verification is carried out by adopting small data:
FIG. 10 is a graph of the trend of predicting the oil temperature of a pipeline by using a BP neural network and an ARMA model when a test data set is adopted, wherein a gray line is a predicted value, a black line is a true value, and it can be seen that most of the true values are covered by the predicted value, and the coincidence ratio of the two values is high. FIG. 11 is a histogram of errors in predicting oil temperature in a pipeline using a BP neural network and an ARMA model with a test data set, and it can be seen that the errors are concentrated between-0.2 and 0.2, most are concentrated at 0 and its vicinity, and the explanation error is small.
FIG. 12 is a graph of the trend of predicting the oil temperature of a pipeline by using a BP neural network and an ARMA model when a training data set is adopted, wherein a gray line is a predicted value, a black line is a true value, and the coincidence ratio of the two values is very high. FIG. 13 is a histogram of errors in predicting oil temperature in a pipeline using a BP neural network and an ARMA model with a training data set, and it can be seen that the errors are concentrated between-0.2 and 0.3, most of the errors are concentrated at 0 and its vicinity, and the explanation error is small.
FIG. 14 is a graph of the trend of the error of predicting the oil temperature of the pipeline by using the BP neural network and the ARMA model when a training data set is adopted, and it can be seen that the error is basically below 0.5. FIG. 15 is a graph of the trend of predicting the oil temperature of the pipeline by using the BP neural network and the ARMA model when a test data set is adopted, and it can be seen that the errors are below 0.5.
The model verified sample data for this time consisted of 4298, of which 80% of the data (3515) were used as the training data set and 20% of the data (783) were used as the test data set, as shown in table 2.
TABLE 2
As can be seen from the data in FIGS. 10-15 and Table 2, the true value and the predicted value of the training dataset have consistent trend change and high contact ratio, and the absolute value of the error between the predicted value and the true value of the oil temperature is less than 0.5 ℃ and reaches 99.91%; and predicting the test data set by using a model obtained by the training data set, wherein the absolute value of the error is less than 0.5 ℃ and reaches 100%. According to the verification of the small amount of data, the oil temperature can be accurately predicted by the ARMA optimized BP neural network oil temperature prediction model, and the oil temperature prediction model has high generalization performance.
BP neural network and ARMA model verification is carried out by adopting a large amount of data:
and (3) downloading relevant data from the first station 1 to the hot station 2, from the hot station 3 to the last station 4 in 2015 in 3 to 2017 in 6 months by using a SCADA system database of a hot oil pipeline, wherein the data amount reaches 39600 pieces. The oil temperatures of the 2# hot station and the 4# end station are predicted by using the BP neural network and the ARMA model, and the results are shown in fig. 16 to 23 and table 3. Fig. 16 is a graph of the predicted and true value trends in the 2# hot station training data set, where the gray bars are predicted values and the black bars are true values, and it can be seen that the two are substantially completely coincident. Fig. 17 is a comparison graph of the error of the 2# hot station training, and it can be seen that the error is substantially below 0.5 (dashed line in the graph). Fig. 18 is a graph of predicted and true value trends for the 2# hot station test data set, where the gray lines are predicted values and the black lines are true values, and it can be seen that the coincidence ratio is high. Fig. 19 is a comparison graph of the test error of # 2 hot station, and it can be seen that most of the errors are much less than 0.5. Fig. 20 is a graph of predicted and true value trends for the 4# end station training data set, where the gray bars are predicted values and the black bars are true values, and it can be seen that the two are substantially completely coincident. Fig. 21 is a comparison graph of 4# end station training error, which can be seen to be substantially below 0.5. FIG. 22 is a graph of predicted and true value trends for the 4# end station test data set, where the gray lines are predicted values and the black lines are true values, and it can be seen that the two are substantially completely coincident. Fig. 23 is a comparison graph of the 4# end station error, and it can be seen that most of the errors are much less than 0.5.
The data in the last two years is adopted, the training and testing sample amount is large, and the ratio of the training data set to the testing data set is still 8: 2. As can be seen from the data in FIGS. 16-23 and Table 3, the temperature of the downstream inbound oil is predicted by using the BP neural network and the ARMA model, the true value is basically coincident with the predicted value, particularly, the prediction error and the true value of the test data set of the two pipe sections have fluctuation except for individual data, the absolute error is less than 0.5 ℃, the data is 99.8 percent, most errors are below 0.2 ℃, and the prediction precision is highly consistent.
TABLE 3
Therefore, the BP neural network and the ARMA model provided by the embodiment of the invention have better actual production and application effects and stronger adaptability and generalization performance.
Fig. 24 is a schematic structural diagram of an oil temperature prediction device according to an embodiment of the present invention. Referring to fig. 24, the apparatus includes: an acquisition module 301, a first prediction module 302, a calculation module 303, and a second prediction module 305.
The obtaining module 301 is configured to obtain a training data set, where the training data set includes a plurality of data samples, and each data sample includes time, historical upstream oil temperature, historical upstream ground temperature, historical downstream arrival ground temperature, historical output, and historical downstream arrival oil temperature; the first prediction module 302 is configured to call an oil temperature prediction model to perform oil temperature prediction based on historical upstream oil temperature, historical upstream ground temperature, historical downstream station-entering ground temperature and historical output in a plurality of data samples to obtain a plurality of predicted historical oil temperatures, the plurality of data samples are continuously extracted within a period of time, and the oil temperature prediction model is obtained by training a back propagation neural network by using a training data set; the calculation module 303 is configured to subtract the plurality of predicted historical oil temperatures from the historical downstream inbound oil temperatures of the plurality of predicted historical oil temperatures, respectively, to obtain a residual value sequence; the second prediction module 304 is configured to invoke an autoregressive moving average model to predict a residual value of the residual value sequence at a next time; the first prediction module 302 is further configured to call an oil temperature prediction model to perform oil temperature prediction based on oil temperature data to be predicted to obtain predicted oil temperature at the next time, where the oil temperature data to be predicted includes upstream oil temperature, upstream ground temperature, downstream arrival ground temperature, and output; the calculating module 303 is further configured to add the residual value at the next time and the predicted oil temperature at the next time to obtain a predicted value of the inbound oil temperature at the next time.
In an implementation manner of the embodiment of the present invention, the obtaining module 301 is configured to extract historical oil temperature data at predetermined time intervals; filtering the extracted historical oil temperature data, and removing abnormal data; and converting the filtered historical oil temperature data into a preset format to obtain a plurality of data samples.
In an implementation manner of the embodiment of the present invention, the second prediction module 304 is further configured to perform denoising processing on the residual value sequence before the autoregressive moving average model is called to predict the residual value at the next time of the residual value sequence.
In an implementation manner of the embodiment of the present invention, the apparatus further includes: and the processing module 305 is configured to use the predicted inbound oil temperature value at the next time as the predicted inbound oil temperature value after a time delay at the next time.
Fig. 25 is a block diagram of a configuration of an oil temperature prediction device according to an embodiment of the present disclosure. The oil temperature prediction device may be a server. Referring to fig. 25, a server 600 includes a Central Processing Unit (CPU)601, a system memory 604 including a Random Access Memory (RAM)602 and a Read Only Memory (ROM)603, and a system bus 605 connecting the system memory 604 and the central processing unit 601. The server 600 also includes a basic input/output system (I/O system) 606, which facilitates the transfer of information between devices within the computer, and a mass storage device 607, which stores an operating system 613, application programs 614, and other program modules 615.
The basic input/output system 606 includes a display 608 for displaying information and an input device 609 such as a mouse, keyboard, etc. for a user to input information. Wherein the display 608 and the input device 609 are connected to the central processing unit 601 through an input output controller 610 connected to the system bus 605. The basic input/output system 606 may also include an input/output controller 610 for receiving and processing input from a number of other devices, such as a keyboard, mouse, or electronic stylus. Similarly, input/output controller 610 also provides output to a display screen, a printer, or other type of output device.
The mass storage device 607 is connected to the central processing unit 601 through a mass storage controller (not shown) connected to the system bus 605. The mass storage device 607 and its associated computer-readable media provide non-volatile storage for the server 600. That is, the mass storage device 607 may include a computer-readable medium (not shown) such as a hard disk or CD-ROM drive.
Without loss of generality, the computer-readable media may comprise computer storage media and communication media. Computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media includes RAM, ROM, EPROM, EEPROM, flash memory or other solid state memory technology, CD-ROM, DVD, or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices. Of course, those skilled in the art will appreciate that the computer storage media is not limited to the foregoing. The system memory 604 and mass storage 607 described above may be collectively referred to as memory.
The server 600 may also operate as a remote computer connected to a network through a network, such as the internet, in accordance with various embodiments of the present invention. That is, the server 600 may be connected to the network 612 through the network interface unit 611 connected to the system bus 605, or may be connected to other types of networks or remote computer systems (not shown) using the network interface unit 611.
The memory further includes one or more programs, the one or more programs are stored in the memory, and the central processing unit 601 implements the oil temperature prediction method shown in fig. 1 or 2 by executing the one or more programs.
In an exemplary embodiment, a non-transitory computer readable storage medium, such as a memory, including instructions executable by a processor of a server to perform the oil temperature prediction methods shown in the various embodiments of the present invention is also provided. For example, the non-transitory computer readable storage medium may be a ROM, a Random Access Memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and should not be taken as limiting the scope of the present invention, which is intended to cover any modifications, equivalents, improvements, etc. within the spirit and scope of the present invention.
Claims (10)
1. A method of predicting oil temperature, the method comprising:
acquiring a training data set, wherein the training data set comprises a plurality of data samples, and each data sample comprises time, historical upstream oil temperature, historical upstream ground temperature, historical downstream station entering ground temperature, historical output and historical downstream station entering oil temperature;
based on historical upstream oil temperature, historical upstream ground temperature, historical downstream station entering ground temperature and historical output in a plurality of data samples, calling an oil temperature prediction model to predict oil temperature to obtain a plurality of predicted historical oil temperatures, wherein the plurality of data samples are continuously extracted within a period of time, the oil temperature prediction model is obtained by training a back propagation neural network by adopting the training data set, and the back propagation neural network is built according to the following method: preliminarily analyzing historical oil temperature data, judging the degree of correlation of the characteristic data to the integral regression algorithm, then building a neural network model, and optimizing parameters of the built algorithm by using a neural network prediction result;
subtracting the plurality of predicted historical oil temperatures from historical downstream incoming oil temperatures of the plurality of predicted historical oil temperatures respectively to obtain a residual value sequence;
calling an autoregressive moving average model to predict a residual value of the residual value sequence at the next moment; the autoregressive moving average model is as follows:
e t ＝f(e t-1 ,e t-2 ,…,e t-n )+ε t
in the formula: f is a linear function determined by an autoregressive moving average model, ε t As random error, e t-1 ,e t-2 ,…,e t-n Is a sequence of residual values, e t Is the residual value at the next moment;
calling the oil temperature prediction model based on oil temperature data to be predicted to perform oil temperature prediction to obtain predicted oil temperature at the next moment, wherein the oil temperature data to be predicted comprises upstream oil temperature, upstream ground temperature, downstream station-entering ground temperature and output;
and adding the residual value at the next moment and the predicted oil temperature at the next moment to obtain a predicted value of the inbound oil temperature at the next moment.
2. The method of claim 1, wherein the obtaining a training data set comprises:
extracting historical oil temperature data at preset time intervals;
filtering the extracted historical oil temperature data to remove abnormal data;
and converting the filtered historical oil temperature data into a preset format to obtain a plurality of data samples.
3. The method according to claim 1 or 2, wherein before the calling the autoregressive moving average model to predict the residual value of the sequence of residual values at the next moment, the method further comprises:
and carrying out noise reduction processing on the residual value sequence.
4. The method according to claim 1 or 2, characterized in that the method further comprises:
and taking the predicted value of the inbound oil temperature at the next moment as the predicted value of the inbound oil temperature delayed for a period of time at the next moment.
5. An oil temperature prediction device, characterized in that the device comprises:
the system comprises an acquisition module, a data acquisition module and a data processing module, wherein the acquisition module is used for acquiring a training data set, the training data set comprises a plurality of data samples, and each data sample comprises time, historical upstream oil temperature, historical upstream ground temperature, historical downstream arrival ground temperature, historical output and historical downstream arrival oil temperature;
the first prediction module is used for calling an oil temperature prediction model to predict oil temperature based on historical upstream oil temperature, historical upstream ground temperature, historical downstream station entering ground temperature and historical output in a plurality of data samples to obtain a plurality of predicted historical oil temperatures, the plurality of data samples are continuously extracted within a period of time, the oil temperature prediction model is obtained by training a back propagation neural network by adopting the training data set, and the back propagation neural network is built according to the following mode: preliminarily analyzing historical oil temperature data, judging the degree of correlation of the characteristic data to the integral regression algorithm, then building a neural network model, and carrying out parameter optimization on the built algorithm by using a neural network prediction result;
the calculation module is used for subtracting the plurality of predicted historical oil temperatures from the historical downstream incoming oil temperatures of the plurality of predicted historical oil temperatures respectively to obtain a residual value sequence;
the second prediction module is used for calling an autoregressive moving average model to predict a residual value of the residual value sequence at the next moment; the autoregressive moving average model is as follows:
e t ＝f(e t-1 ,e t-2 ,…,e t-n )+ε t
in the formula: f is a linear function determined by an autoregressive moving average model, ε t For random error, e t-1 ,e t-2 ,…,e t-n As a sequence of residual values, e t Is the residual value at the next moment;
the first prediction module is also used for calling the oil temperature prediction model to perform oil temperature prediction based on oil temperature data to be predicted to obtain predicted oil temperature at the next moment, and the oil temperature data to be predicted comprises upstream oil temperature, upstream geothermal temperature, downstream station-entering geothermal temperature and output;
the calculation module is further configured to add the residual value at the next time and the predicted oil temperature at the next time to obtain a predicted value of the inbound oil temperature at the next time.
6. The device according to claim 5, wherein the obtaining module is configured to extract historical oil temperature data at predetermined time intervals; filtering the extracted historical oil temperature data, and removing abnormal data; and converting the filtered historical oil temperature data into a preset format to obtain a plurality of data samples.
7. The apparatus according to claim 5 or 6, wherein the second prediction module is further configured to perform denoising on the residual value sequence before the invoking of the auto-regressive moving average model predicts the residual value at the next time of the residual value sequence.
8. The apparatus of claim 5 or 6, further comprising: and the processing module is used for taking the predicted value of the incoming oil temperature at the next moment as the predicted value of the incoming oil temperature delayed for a period of time at the next moment.
9. An oil temperature prediction device, characterized in that the device comprises: a processor; a memory for storing processor-executable instructions; wherein the processor is configured to execute executable instructions stored in the memory to implement the oil temperature prediction method of any one of claims 1-4.
10. A computer-readable storage medium, wherein instructions in the computer-readable storage medium, when executed by a processor of an oil temperature prediction device, enable the oil temperature prediction device to perform the oil temperature prediction method of any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910057412.1A CN111539545B (en) | 2019-01-22 | 2019-01-22 | Oil temperature prediction method and device and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910057412.1A CN111539545B (en) | 2019-01-22 | 2019-01-22 | Oil temperature prediction method and device and computer readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111539545A CN111539545A (en) | 2020-08-14 |
| CN111539545B true CN111539545B (en) | 2022-08-30 |
Family
ID=71976640
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910057412.1A Active CN111539545B (en) | 2019-01-22 | 2019-01-22 | Oil temperature prediction method and device and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111539545B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113420202B (en) * | 2021-07-15 | 2024-04-02 | 上海明略人工智能(集团)有限公司 | Method and device for predicting keyword search times, electronic equipment and storage medium |
| CN116703037B (en) * | 2023-07-06 | 2023-12-19 | 中建二局第二建筑工程有限公司 | Monitoring method and device based on road construction |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010236571A (en) * | 2009-03-30 | 2010-10-21 | Toyota Motor Corp | Method of calculating oil temperature of automatic transmission |
| CN109190314A (en) * | 2018-10-30 | 2019-01-11 | 河海大学 | The prediction technique of the power transformer top-oil temperature of neural network based on Adam optimization |
| CN109242147A (en) * | 2018-08-07 | 2019-01-18 | 重庆大学 | Signal fused fan condition prediction technique based on Bp neural network |
-
2019
- 2019-01-22 CN CN201910057412.1A patent/CN111539545B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2010236571A (en) * | 2009-03-30 | 2010-10-21 | Toyota Motor Corp | Method of calculating oil temperature of automatic transmission |
| CN109242147A (en) * | 2018-08-07 | 2019-01-18 | 重庆大学 | Signal fused fan condition prediction technique based on Bp neural network |
| CN109190314A (en) * | 2018-10-30 | 2019-01-11 | 河海大学 | The prediction technique of the power transformer top-oil temperature of neural network based on Adam optimization |
Non-Patent Citations (1)
| Title |
|---|
| "长呼原油管道油温控制优化运行";田志强等;《化工管理》;20141120;193-194 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111539545A (en) | 2020-08-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111539545B (en) | Oil temperature prediction method and device and computer readable storage medium | |
| Kapoor et al. | Weather forecasting using sliding window algorithm | |
| CN112131782B (en) | Multi-loop intelligent factory edge side digital twin scene coupling device | |
| CN111625516B (en) | Method, apparatus, computer device and storage medium for detecting data state | |
| CN110457184B (en) | Chemical engineering abnormal cause and effect analysis and graph display method based on time sequence fluctuation correlation | |
| CN111309539A (en) | Abnormity monitoring method and device and electronic equipment | |
| CN110210660B (en) | Ultra-short-term wind speed prediction method | |
| CN111275255A (en) | Construction method of concrete dam deformation monitoring and forecasting model | |
| Van der Pol et al. | Optimal dike investments under uncertainty and learning about increasing water levels | |
| CN112085285B (en) | Bus load prediction method, device, computer equipment and storage medium | |
| CN106099932B (en) | Day-ahead planning power flow analysis method considering uncertainty time-space correlation | |
| CN113868953A (en) | Multi-unit operation optimization method, device and system in industrial system and storage medium | |
| CN110059126B (en) | LKJ abnormal value data-based complex correlation network analysis method and system | |
| JP2019125306A (en) | Data processing method, data processing device and program | |
| CN112949201B (en) | Wind speed prediction method and device, electronic equipment and storage medium | |
| CN105787283B (en) | A kind of earthen ruins monitoring data amendment approximating method based on temporal correlation | |
| CN105389637A (en) | Effective workday construction method for monthly electricity consumption prediction model | |
| CN114897224A (en) | Urban carbon emission prediction method based on mathematical model | |
| Christer et al. | A recent mathematical development in maintenance theory | |
| Wang et al. | Degradation prediction method by use of autoregressive algorithm | |
| Lee et al. | Very short-Term wind power ensemble forecasting without numerical weather prediction through the predictor design | |
| CN107844679A (en) | A kind of segmented blast furnace molten iron silicon content Forecasting Methodology and device | |
| Béguinot | On General Mathematical Constraints Applying to the Kinetics of Species Discovery during Progressive Sampling: Consequences on the Theoretical Expression of the Species Accumulation Curve | |
| CN115809732A (en) | Distributed photovoltaic power generation power prediction method and system | |
| CN110244563B (en) | Neural network internal model controller model mismatch identification and online updating method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |