CN117592571B - Air conditioning unit fault type diagnosis method and system based on big data - Google Patents
Air conditioning unit fault type diagnosis method and system based on big data Download PDFInfo
- Publication number
- CN117592571B CN117592571B CN202311656595.1A CN202311656595A CN117592571B CN 117592571 B CN117592571 B CN 117592571B CN 202311656595 A CN202311656595 A CN 202311656595A CN 117592571 B CN117592571 B CN 117592571B
- Authority
- CN
- China
- Prior art keywords
- sequence
- feature
- base sequences
- sampling
- air conditioning
- 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
- 238000004378 air conditioning Methods 0.000 title claims abstract description 117
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000003745 diagnosis Methods 0.000 title claims abstract description 42
- 238000005070 sampling Methods 0.000 claims abstract description 98
- 238000012545 processing Methods 0.000 claims description 34
- 238000012549 training Methods 0.000 claims description 23
- 230000007246 mechanism Effects 0.000 claims description 21
- 238000003672 processing method Methods 0.000 claims description 20
- 230000008569 process Effects 0.000 claims description 17
- 238000010606 normalization Methods 0.000 claims description 11
- 238000004590 computer program Methods 0.000 claims description 10
- 230000006870 function Effects 0.000 claims description 4
- 238000004891 communication Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000009833 condensation Methods 0.000 description 3
- 230000005494 condensation Effects 0.000 description 3
- 238000012544 monitoring process Methods 0.000 description 3
- 239000003507 refrigerant Substances 0.000 description 3
- 238000005057 refrigeration Methods 0.000 description 3
- 230000032683 aging Effects 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000017525 heat dissipation Effects 0.000 description 2
- 238000003058 natural language processing Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007797 corrosion Effects 0.000 description 1
- 238000005260 corrosion Methods 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000002405 diagnostic procedure Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000011159 matrix material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 238000013021 overheating Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000005096 rolling process Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 230000001502 supplementing effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 238000010977 unit operation Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Computational Linguistics (AREA)
- Computing Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The application relates to the technical field of air conditioner fault diagnosis, and provides an air conditioner unit fault type diagnosis method and system based on big data, wherein the method comprises the following steps: acquiring a plurality of sampling base sequences and reasoning base sequences corresponding to the current moment; the inference base sequence includes: a past known feature sequence, a future known first feature sequence, and a future unknown second feature sequence; inputting the plurality of sampling base sequences and the inference base sequences into a feature sequence prediction model to obtain predicted values of all features in a second feature sequence in the inference base sequences; and diagnosing the air conditioning unit based on the predicted value of each feature in the second feature sequence to obtain the fault type of the air conditioning unit. The embodiment of the application can also realize the estimation of the parameters of the air conditioning unit under the actual working condition without the support of a physical model, reduces the estimation error of the parameters of the air conditioning unit under the actual working condition, and improves the accuracy of fault type diagnosis of the air conditioning unit.
Description
Technical Field
The application relates to the technical field of air conditioner fault diagnosis, in particular to an air conditioner unit fault type diagnosis method and system based on big data.
Background
In the fault diagnosis process of the air conditioning unit, the fault type diagnosis of the air conditioning unit is mainly carried out based on a physical model, so that the accurate fault diagnosis, the early prediction of potential faults, the remote monitoring and diagnosis and the like can be carried out by utilizing historical data and real-time monitoring data based on the physical model. However, the physical model may be subject to false alarms or false judgments, and particularly under complex and variable working conditions, the algorithm may not accurately judge the fault type, resulting in erroneous maintenance decisions.
Disclosure of Invention
The application aims to provide a fault type diagnosis method and system for an air conditioning unit based on big data, which are used for realizing the prediction of the parameters of the air conditioning unit under the actual working condition under the condition that no physical model support is needed, reducing the estimation error of the parameters of the air conditioning unit under the actual working condition and improving the accuracy of the fault type diagnosis of the air conditioning unit.
In a first aspect, the present application provides a fault type diagnosis method for an air conditioning unit based on big data, including:
Acquiring a plurality of sampling base sequences and reasoning base sequences corresponding to the current moment; the inference base sequence includes: a past known feature sequence, a future known first feature sequence, and a future unknown second feature sequence; the second characteristic sequence is a characteristic sequence to be inferred;
Inputting the plurality of sampling base sequences and the reasoning base sequence into a feature sequence prediction model to obtain predicted values of all features in the second feature sequence; the predicted value is used for representing component state information of the air conditioning unit;
Diagnosing the air conditioning unit based on the predicted value of each feature in the second feature sequence to obtain the fault type of the air conditioning unit;
the plurality of sampling base sequences are base sequences which have been sampled before the current moment; the characteristic sequence prediction model is constructed based on a self-attention mechanism model; the feature sequence prediction model comprises: an inference encoder, a sampling encoder and a decoder; the sampling encoder is used for processing the plurality of sampling base sequences to obtain first result data; the inference encoder is used for processing the past known characteristic sequences to obtain second result data; the decoder is used for decoding the first result data, the second result data and the first feature sequence to obtain predicted values of all features in the second feature sequence; the base sequence comprises: and the component state information is collected in the running process of the air conditioning unit.
In a second aspect, the present application also provides an air conditioning unit fault type diagnosis system based on big data, including:
The acquisition module is used for acquiring a plurality of sampling base sequences and reasoning base sequences corresponding to the current moment; the inference base sequence includes: a past known feature sequence, a future known first feature sequence, and a future unknown second feature sequence; the second characteristic sequence is a characteristic sequence to be inferred;
the prediction module is used for inputting the plurality of sampling base sequences and the reasoning base sequences into a feature sequence prediction model to obtain predicted values of all features in the second feature sequence; the predicted value is used for representing component state information of the air conditioning unit;
The fault type diagnosis module is used for diagnosing the air conditioning unit based on the predicted value of each feature in the second feature sequence to obtain the fault type of the air conditioning unit;
the plurality of sampling base sequences are base sequences which have been sampled before the current moment; the characteristic sequence prediction model is constructed based on a self-attention mechanism model; the feature sequence prediction model comprises: an inference encoder, a sampling encoder and a decoder; the sampling encoder is used for processing the plurality of sampling base sequences to obtain first result data; the inference encoder is used for processing the past known characteristic sequences to obtain second result data; the decoder is used for decoding the first result data, the second result data and the first feature sequence to obtain predicted values of all features in the second feature sequence; the base sequence comprises: and the component state information is collected in the running process of the air conditioning unit.
In a third aspect, the present application also provides a computer program product comprising a computer program which, when executed by a processor, implements the big data based air conditioning unit fault type diagnosis method as described in any of the above.
In a fourth aspect, the present application further provides an electronic device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, where the processor implements the method for diagnosing a fault type of an air conditioning unit based on big data as described in any one of the above when executing the program.
In a fifth aspect, the present application also provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the big data based air conditioning unit fault type diagnosis method as described in any of the above.
According to the fault type diagnosis method and system for the air conditioning unit based on the big data, provided by the application, under the condition that the support of a physical model is not needed, the estimation of the parameters of the air conditioning unit under the actual working condition can be realized, the estimation error of the parameters of the air conditioning unit under the actual working condition is reduced, and the accuracy of the fault type diagnosis of the air conditioning unit is improved.
Drawings
In order to more clearly illustrate the application or the technical solutions of the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described, and it is obvious that the drawings in the description below are some embodiments of the application, and other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic flow chart of a fault type diagnosis method of an air conditioning unit based on big data;
FIG. 2 is a schematic representation of the generation of the present application using base sequences;
FIG. 3 is a schematic representation of the generation of an inference base sequence provided by the present application;
FIG. 4 is a schematic structural diagram of a feature sequence prediction model provided by the application;
FIG. 5 is a schematic diagram of the structure of the fault type diagnosis system of the air conditioning unit based on big data provided by the application;
fig. 6 is a schematic structural diagram of an electronic device provided by the present application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the technical solutions of the present application will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present application, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
The terms first, second and the like in the description and in the claims, are used for distinguishing between similar elements and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged, as appropriate, such that embodiments of the present application may be implemented in sequences other than those illustrated or described herein, and that the objects identified by "first," "second," etc. are generally of a type, and are not limited to the number of objects, such as the first object may be one or more. Furthermore, in the description and claims, "and/or" means at least one of the connected objects, and the character "/", generally means that the associated object is an "or" relationship.
The following description is made with respect to terms related to embodiments of the present application:
The Self-Attention mechanism model in the embodiment of the application is a transducer model, which is a neural network model based on Self-Attention mechanism (Self-Attention), and is originally used for machine translation in natural language processing tasks. Compared with a traditional sequence model, such as a cyclic neural network (Recurrent Neural Network, RNN), the transducer is more suitable for processing long-distance dependence and can perform parallel calculation, so that the model training speed is increased.
In the transducer, K, V, Q are the results of three linear transformations (matrix multiplications), representing representations of keys (keys), values (values), and queries (Query), respectively. Bond (K): for encoding the input sequence, plays a role in the self-attention mechanism in marking the importance of each position in the input. Value (V): for storing information for each position in the input sequence as a response to the query. Query (Q): for obtaining information related to the current position in the input sequence. The attention weight is obtained by calculating the similarity between the query and the key, and the weighted sum of the weight and the value is carried out to obtain the context representation of each query position. Such a self-attention mechanism enables the model to build associations between different locations and focus on important parts of the input sequence.
In the transducer, feature extraction and representation learning is performed by a multi-layered stack of self-attention layers and feedforward neural network layers, thereby completing the task. Through a large amount of training data and a self-adaptive mechanism, the transducer can learn the semantic representation of the input sequence, and the breakthrough performance in the fields of natural language processing and the like is realized.
The self-attention mechanism model in the embodiment of the application is constructed based on a transducer structure. On the one hand, the attention mechanism of the transducer can implicitly extract the global information of the sampling base sequence and combine the global information with the reasoning base sequence to infer a required characteristic sequence; on the other hand, the attention mechanism can capture internal cross-feature and cross-length internal relations in the base sequence, so that the model can learn the mapping relation between feature sequences. The multi-headed attention of the transducer enables parallel computation of the model.
When predicting the state parameters of the air conditioning unit, the physical model used by the air conditioning unit state prediction method based on the physical model or rule is often from laboratory data modeling, and has larger difference from the actual air conditioning unit running condition. The traditional fault type diagnosis model of the air conditioning unit based on big data based on deep learning can only predict the future state of the air conditioning unit according to a small section of historical information in the past at the current moment, and neglects the state information implied by accumulated data of the air conditioning unit in long-time operation.
Aiming at the technical problems in the related art, the embodiment of the application provides a self-attention mechanism model. The model in the embodiment of the application is a data-driven method, and does not need a physical model of an air conditioning unit.
The fault type diagnosis method of the air conditioning unit based on big data provided by the embodiment of the application is described in detail through specific embodiments and application scenes thereof with reference to the accompanying drawings.
As shown in fig. 1, the method for diagnosing a fault type of an air conditioning unit based on big data according to the embodiment of the present application may include the following steps 101 and 103:
and 101, acquiring a plurality of sampling base sequences and reasoning base sequences corresponding to the current moment.
Wherein the inference base sequence comprises: a past known feature sequence, a future known first feature sequence, and a future unknown second feature sequence.
The second feature sequence is a feature sequence to be inferred.
The model in the embodiment of the application uses the global history information and the current history information of the air conditioning unit to input data, and a certain preprocessing is needed before the data is input into the model. In air conditioning unit fault prediction planning and control, a Frenet coordinate system based on time s is often used, and therefore, the above-mentioned base sequences (including sampling base sequences and inference base sequences) are also sequence data constructed based on the Frenet coordinate system.
It will be appreciated that, as shown in fig. 2, assuming that the number of features of the air conditioning unit used is m and the number of data points of the sequence is l in total, each piece of data rl×m may be referred to as a base sequence. If the interval between two adjacent points in the sequence data is delta s, one base sequence covers air conditioning unit data with time of l x delta s in a Frenet coordinate system. In the running process of the air conditioning unit, the high-frequency data of the air conditioning unit characteristic is collected based on a controller area network bus (Controller Area Network, CAN) and stored in a memory. Based on the sliding window samples, a base sequence can be generated at each interval deltas. It should be noted that, in the time corresponding to the sliding window, the air conditioning unit does not perform feature sampling.
It will be appreciated that, among the base sequences that have been generated, n base sequences may be selected as the air conditioning unit operation history information, i.e., the plurality of sampling base sequences described above. The model can learn the characteristics which are not changed greatly for a long time in the running process, such as weather conditions, wind resistance, rolling resistance and other information, through the sampling base sequence, and the model is similar to providing the global characteristics of the air conditioning unit. As shown in fig. 2, the sampling base sequence is selected using a sliding window with a spacing Ls. Each set of sample basis sequences remains unchanged over a longer distance.
At the present moment, the base sequence is incomplete and consists of three parts, namely a past known characteristic sequence, a future known characteristic sequence (namely the first characteristic sequence known in the future) and a characteristic sequence to be inferred in the future (namely the second characteristic sequence unknown in the future), and the structure of the base sequence is shown in figure 3. The feature sequences known in the past are a group ofIs the data from the current time to the past lp length. The future known features are some number m k of features, which are known from the current time to the future lf length (i=lf+lp), such as altitude, air conditioning unit speed sequence (given by the air conditioning unit fault prediction planning module). The future features to be inferred are air conditioning unit feature sequences which need to be inferred, the length is lf, and the feature quantity is m u(m＝mk+mu). This part is filled in by zero values before no inference results are obtained. These unknown features require modeling to be derived by inference in combination with past and future known feature sequences, supplementing the base sequence with integrity. The model is required to supplement the complete base sequence by reasoning at the current moment is called a reasoning base sequence.
At each current time, as can be seen from the above, there are n sampling base sequences of rl×m and 1 inference base sequence of rl×m, which together form input data of R (n+1) ×l×m. After model reasoning, obtainOutput results of (2).
Specifically, based on the above description, the above step 101 may include the following steps 101a1 and 101a2:
Step 101a1, acquiring standard base sequences acquired at a plurality of moments before the current moment, and obtaining a plurality of sampling base sequences.
Step 101a2, obtaining the past known feature sequence in the inference base sequence which is already acquired before the current moment, and constructing the inference base sequence based on each known feature in the standard base sequence and each unknown feature in the standard base sequence after the current moment.
The standard base sequence comprises feature data corresponding to a plurality of operation stages, and the feature quantity of the feature data corresponding to each stage is a first feature quantity; each feature in the second sequence of features is filled with a zero value.
And 102, inputting the plurality of sampling base sequences and the reasoning base sequences into a feature sequence prediction model to obtain predicted values of all features in the second feature sequence.
The predicted values are used for representing component state information of the air conditioning unit, and the plurality of sampling base sequences are base sequences which have been sampled before the current moment; the characteristic sequence prediction model is constructed based on a self-attention mechanism model; the feature sequence prediction model comprises: an inference encoder, a sampling encoder and a decoder; the sampling encoder is used for processing the plurality of sampling base sequences to obtain first result data; the inference encoder is used for processing the past known characteristic sequences to obtain second result data; the decoder is used for decoding the first result data, the second result data and the first feature sequence to obtain predicted values of all features in the second feature sequence; the base sequence comprises: component state information collected in the running process of the air conditioning unit.
It should be noted that, in the embodiment of the present application, a complete sequence is called a base sequence, and a feature sequence is a subsequence of the base sequence.
Illustratively, as shown in fig. 4, the above-described feature sequence prediction model is mainly composed of 2 encoders Transformer Encoder and 1 decoder Transformer Decoder. Each Transformer Encoder is a stack of multiple encoder layers Transformer Encoder Layer, the Encoder input to the sample base sequence is referred to as a sample encoder, and the input to the inference base sequence is referred to as an inference encoder. Transformer Decoder is a stack of multiple decoder layers Transformer Decoder Layer, referred to as a decoder.
Specifically, the step 102 may include the following steps 102a, 102b, and 102c:
102a, processing the plurality of sampling base sequences according to a first preset processing method, and inputting the processed plurality of sampling base sequences into the sampling encoder to obtain first result data.
Step 102b, after the feature sequence known in the past is processed according to a second preset processing method, the feature sequence is input into the inference encoder, and second result data is obtained.
Step 102c, processing the first feature sequence according to the second preset processing method to obtain third result data.
The first preset processing method is used for reducing the number dimension and the feature dimension of the sampling base sequence; the second preset processing method is used for reducing the feature dimension of the feature sequence; the sampling encoder and the inference encoder are comprised of a first number of encoder layers; the decoder is made up of a second number of decoder layers.
The first and second amounts may be the same or different, for example.
It should be noted that, the steps 102a, 102b, and 102c may be performed simultaneously or sequentially.
Specifically, the first preset processing method and the second preset processing method in the above steps may be performed according to the following steps, and the above step 102a may further include the following steps 102a1 and 102a2:
step 102a1, reducing the number dimension of the plurality of sampling base sequences to a preset number dimension through a full connection layer, and obtaining a first low-dimension data.
Step 102a2, reducing the feature dimension of the first low-dimensional data to a preset feature dimension through a full connection and position coding layer to obtain second low-dimensional data, and adding position coding information to the second low-dimensional data to obtain the first result data.
Specifically, the step 102b may further include the following step 102b1:
Step 102b1, reducing the feature dimension of the past known feature sequence to a preset feature dimension through a full connection and position coding layer to obtain third low-dimensional data, and adding position coding information to the third low-dimensional data to obtain the second result data.
Specifically, the step 102c may further include the following step 102c1:
And 102c1, reducing the feature dimension of the first feature sequence to a preset feature dimension through a full connection and position coding layer to obtain fourth low-dimensional data, and adding position coding information to the fourth low-dimensional data to obtain third result data.
Illustratively, each base sequence needs to undergo full-join layer dimension conversion (i.e., feature dimension m decreases to h, which is the model dimension) before entering the model. For the sample base sequence, the number dimension needs to be reduced from n to 1 through an additional full connection layer. Because the transducer can not sense the relation between the front and the back of the sequence, and position codes are added to the input data, sinusoidal position codes can be adopted in the embodiment of the application, namely
PE(pos，2i)＝sin(pos/100002i/h)
PE(pos，2i+1)＝cos(pos/100002i/h)
Wherein PE (pos, 2 i) and PE (pos, 2i+1) are position codes, pos is the position of a data point, and i is a characteristic dimension.
The results of the inference encoder and the sampling encoder are spliced and then input to the decoder as K, V. Future known sequences are subjected to full-concatenated layer and position coding and then serve as Q input decoders. The result of the decoder is a sliceThe data of (2) is also required to be subjected to full connection layer to realize dimension reduction and become a strip
The sequence of the desired features obtained is inferred.
Specifically, the step 102 may further include the following step 102d:
102d, inputting the first result data and the second result data into the decoder as a key K and a value V after being spliced, and inputting the third result data into the decoder as a query Q and the key K and the value V at the same time, and predicting the values of all the features in the second feature sequence to obtain predicted values of all the features in the second feature sequence.
Illustratively, as shown in fig. 4, n sample base sequences (i.e., the above-mentioned plurality of sample base sequences) need to be reduced in number dimension by a fully-connected layer and reduced in feature dimension by a fully-connected layer+position-coding layer (i.e., the above-mentioned fully-connected and position-coding layers) before being input to the sample encoder, and then input to the sample encoder for processing. Similarly, both future known feature sequences and past known feature sequences require feature dimensions to be reduced by a full concatenation layer+position coding layer (i.e., the full concatenation and position coding layers described above).
Illustratively, as shown in fig. 4, the data output by the encoder may also increase the feature dimension from h to m through a fully connected layer. The output inferred future feature sequence may then be added to the inferential base sequence.
Optionally, in the embodiment of the present application, a model may be constructed by using a deep learning framework such as PyTorch, tensorflow, and during training, the operation data of the air conditioning component is processed to obtain a corresponding base sequence, and then the data is processed by using a minimum maximum normalization method, that is
Wherein,For normalized processed data, X is the value of each feature, xmax is the maximum value of each feature, xmin is the minimum value of each feature, the data value is scaled to be within the range of 0-1 through normalization processing, and then the data is input into a model for training. Here the maximum and minimum values of each feature need to be recorded for inverse normalization. During training, a mean square error (Mean Squared Error, MSE) is used as a loss function, so that back propagation and updating of model parameters are realized.
And step 103, diagnosing the air conditioning unit based on the predicted values of all the features in the second feature sequence to obtain the fault type of the air conditioning unit.
Optionally, a predicted value-fault type matching table of the air conditioning unit is obtained, as shown in table 1.
TABLE 1
Compressor failure: the compressor is one of the most critical components in an air conditioning unit, and common faults include difficult start-up, abnormal operation, overheating, and the like. Condenser problem: condensers are important components for heat dissipation, and common problems include corrosion, clogging, fan failure, etc., resulting in a decrease in heat dissipation effect. Evaporator failure: the evaporator plays a role in absorbing heat in the refrigeration process, and common problems include poor condensation, blockage and the like. Electrical problems: electrical problems may cause the air conditioning unit to fail to start or to operate abnormally, such as circuit failure, loose wiring, burned fuses, etc. Refrigerant leakage: refrigerant leakage can lead to reduced or even complete failure of the refrigeration effect, possibly due to line damage or seal aging. Sensor failure: the sensor is used for monitoring parameters such as temperature, pressure and the like, and if the sensor is damaged or fails, inaccurate system control can be caused. Fan problem: fans are used to circulate air and dissipate heat, common problems including fan damage, rotor blockage, bearing aging, and the like. Drainage problem: the air conditioning unit can generate condensation in the refrigeration process, and if a drainage system is blocked or a drainage pipeline is damaged, water leakage or condensation can be caused.
Further, diagnosing the air conditioning unit according to the predicted value-fault type matching table and the predicted value of each feature in the second feature sequence to obtain the fault type of the air conditioning unit.
In one embodiment, the predicted value of each feature in the second feature sequence is "0, a, 0", and the fault type of the air conditioning unit is a compressor fault. The predicted value of each feature in the second feature sequence is '0, aa,0, c, 0', and the fault type of the air conditioning unit is refrigerant leakage+evaporator fault.
Illustratively, based on the above description, before step 102, the method for diagnosing a fault type of an air conditioning unit based on big data according to the embodiment of the present application may further include the following steps 104 and 105:
step 104, acquiring a plurality of sample sampling base sequences as training samples to be processed, and carrying out normalization processing on the training samples according to a maximum and minimum normalization method to obtain target training samples.
And 105, training a self-attention mechanism model by using the target training sample to obtain the characteristic sequence prediction model.
In the training process, the mean square error is used as a loss function to realize the reverse propagation and the updating of the model parameters.
Illustratively, upon model deployment, a trained model file is loaded. When the air conditioning unit operates, the sampling base sequences with the number of n need to be recorded and stored at first. The input base sequence can then be processed using the same normalization and inferred using the model each time an inference base sequence is generated. And carrying out inverse normalization on the obtained reasoning result to obtain the required characteristic sequence.
The application divides the air conditioning unit sequence data into a sampling base sequence and an reasoning base sequence. The sampling base sequence can be regarded as a global variable when the air conditioning unit operates, and provides global characteristics of the state of the air conditioning unit. The inference base sequence is a local variable near the current moment, provides the current air conditioning unit state, and the model can infer future characteristics based on the information.
According to the difference between the sampling base sequence and the reasoning base sequence, the model provided by the application uses 2 decoders and 1 encoder to implicitly learn the characteristic association between the sequences. Compared with the defect that the traditional sequence inference model can only infer according to the current information, the model provided by the application can extract the global information implied by the sampling base sequence, and improves the regression accuracy using the inference base sequence.
According to the air conditioning unit fault type diagnosis method based on big data, provided by the embodiment of the application, under the condition that the physical model support is not needed, the estimation of the air conditioning unit parameters under the actual working condition can be realized, the estimation error of the air conditioning unit parameters under the actual working condition is reduced, and the accuracy of the air conditioning unit fault type diagnosis is improved.
In the embodiment of the present application, the method is shown in the drawings. The fault type diagnosis method of the air conditioning unit based on big data is exemplified by a figure in combination with the embodiment of the application. In specific implementation, the fault type diagnosis method of the air conditioning unit based on big data shown in the above method drawings may be further implemented in combination with any other drawing that may be combined and is illustrated in the above embodiment, and will not be described herein.
The fault type diagnosis system of the air conditioning unit based on big data provided by the application is described below, and the fault type diagnosis method of the air conditioning unit based on big data described below and the fault type diagnosis method of the air conditioning unit based on big data described above can be correspondingly referred to each other.
Fig. 5 is a schematic structural diagram of an air conditioning unit fault type diagnosis system based on big data according to an embodiment of the present application, as shown in fig. 5, specifically including:
The acquisition module 501 is configured to acquire a plurality of sampling base sequences and inference base sequences corresponding to a current moment; the inference base sequence includes: a past known feature sequence, a future known first feature sequence, and a future unknown second feature sequence; the second characteristic sequence is a characteristic sequence to be inferred;
the prediction module 502 is configured to input the plurality of sampling base sequences and the inference base sequence into a feature sequence prediction model to obtain predicted values of each feature in the second feature sequence; the predicted value is used for representing component state information of the air conditioning unit;
A fault type diagnosis module 503, configured to diagnose the air conditioning unit based on the predicted values of the features in the second feature sequence, to obtain a fault type of the air conditioning unit;
The plurality of sampling base sequences are base sequences which have been sampled before the current moment; the characteristic sequence prediction model is constructed based on a self-attention mechanism model; the feature sequence prediction model comprises: an inference encoder, a sampling encoder and a decoder; the sampling encoder is used for processing the plurality of sampling base sequences to obtain first result data; the inference encoder is used for processing the past known characteristic sequences to obtain second result data; the decoder is used for decoding the first result data, the second result data and the first feature sequence to obtain predicted values of all features in the second feature sequence; the base sequence comprises: component state information collected in the running process of the air conditioning unit.
According to the fault type diagnosis system for the air conditioning unit based on the big data, provided by the application, under the condition that the support of a physical model is not needed, the prediction of the parameters of the air conditioning unit under the actual working condition can be realized, the estimation error of the parameters of the air conditioning unit under the actual working condition is reduced, and the accuracy of the fault type diagnosis of the air conditioning unit is improved.
Fig. 6 illustrates a physical schematic diagram of an electronic device, as shown in fig. 6, which may include: processor 610, communication interface (Communications Interface) 620, memory 630, and communication bus 640, wherein processor 610, communication interface 620, memory 630 communicate with each other via communication bus 640. The processor 610 may invoke logic instructions in the memory 630 to perform a big data based air conditioning unit fault type diagnostic method comprising:
Acquiring a plurality of sampling base sequences and reasoning base sequences corresponding to the current moment; the inference base sequence includes: a past known feature sequence, a future known first feature sequence, and a future unknown second feature sequence; the second characteristic sequence is a characteristic sequence to be inferred;
Inputting the plurality of sampling base sequences and the reasoning base sequence into a feature sequence prediction model to obtain predicted values of all features in the second feature sequence; the predicted value is used for representing component state information of the air conditioning unit;
Diagnosing the air conditioning unit based on the predicted value of each feature in the second feature sequence to obtain the fault type of the air conditioning unit;
The plurality of sampling base sequences are base sequences which have been sampled before the current moment; the characteristic sequence prediction model is constructed based on a self-attention mechanism model; the feature sequence prediction model comprises: an inference encoder, a sampling encoder and a decoder; the sampling encoder is used for processing the plurality of sampling base sequences to obtain first result data; the inference encoder is used for processing the past known characteristic sequences to obtain second result data; the decoder is used for decoding the first result data, the second result data and the first feature sequence to obtain predicted values of all features in the second feature sequence; the base sequence comprises: and the component state information is collected in the running process of the air conditioning unit.
Further, the logic instructions in the memory 630 may be implemented in the form of software functional units and stored in a computer-readable storage medium when sold or used as a stand-alone product. Based on this understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
In another aspect, the present application also provides a computer program product comprising a computer program stored on a computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, enable the computer to perform the big data based air conditioning unit fault type diagnosis method provided by the above methods, the method comprising:
Acquiring a plurality of sampling base sequences and reasoning base sequences corresponding to the current moment; the inference base sequence includes: a past known feature sequence, a future known first feature sequence, and a future unknown second feature sequence; the second characteristic sequence is a characteristic sequence to be inferred;
Inputting the plurality of sampling base sequences and the reasoning base sequence into a feature sequence prediction model to obtain predicted values of all features in the second feature sequence; the predicted value is used for representing component state information of the air conditioning unit;
Diagnosing the air conditioning unit based on the predicted value of each feature in the second feature sequence to obtain the fault type of the air conditioning unit;
The plurality of sampling base sequences are base sequences which have been sampled before the current moment; the characteristic sequence prediction model is constructed based on a self-attention mechanism model; the feature sequence prediction model comprises: an inference encoder, a sampling encoder and a decoder; the sampling encoder is used for processing the plurality of sampling base sequences to obtain first result data; the inference encoder is used for processing the past known characteristic sequences to obtain second result data; the decoder is used for decoding the first result data, the second result data and the first feature sequence to obtain predicted values of all features in the second feature sequence; the base sequence comprises: and the component state information is collected in the running process of the air conditioning unit.
In still another aspect, the present application also provides a computer-readable storage medium having stored thereon a computer program which is implemented when executed by a processor to perform the above-provided big data based air conditioning unit fault type diagnosis method, the method comprising:
Acquiring a plurality of sampling base sequences and reasoning base sequences corresponding to the current moment; the inference base sequence includes: a past known feature sequence, a future known first feature sequence, and a future unknown second feature sequence; the second characteristic sequence is a characteristic sequence to be inferred;
Inputting the plurality of sampling base sequences and the reasoning base sequence into a feature sequence prediction model to obtain predicted values of all features in the second feature sequence; the predicted value is used for representing component state information of the air conditioning unit;
Diagnosing the air conditioning unit based on the predicted value of each feature in the second feature sequence to obtain the fault type of the air conditioning unit;
The plurality of sampling base sequences are base sequences which have been sampled before the current moment; the characteristic sequence prediction model is constructed based on a self-attention mechanism model; the feature sequence prediction model comprises: an inference encoder, a sampling encoder and a decoder; the sampling encoder is used for processing the plurality of sampling base sequences to obtain first result data; the inference encoder is used for processing the past known characteristic sequences to obtain second result data; the decoder is used for decoding the first result data, the second result data and the first feature sequence to obtain predicted values of all features in the second feature sequence; the base sequence comprises: and the component state information is collected in the running process of the air conditioning unit.
The system embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.
From the above description of the embodiments, it will be apparent to those skilled in the art that the embodiments may be implemented by means of software plus necessary general hardware platforms, or of course may be implemented by means of hardware. Based on this understanding, the foregoing technical solution may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a computer readable storage medium, such as ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method described in the respective embodiments or some parts of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present application, and are not limiting; although the application has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present application.
Claims (7)
1. The fault type diagnosis method of the air conditioning unit based on big data is characterized by comprising the following steps of:
Acquiring a plurality of sampling base sequences and reasoning base sequences corresponding to the current moment; the inference base sequence includes: a past known feature sequence, a future known first feature sequence, and a future unknown second feature sequence; the second characteristic sequence is a characteristic sequence to be inferred;
Inputting the plurality of sampling base sequences and the reasoning base sequence into a feature sequence prediction model to obtain predicted values of all features in the second feature sequence; the predicted value is used for representing component state information of the air conditioning unit;
Diagnosing the air conditioning unit based on the predicted value of each feature in the second feature sequence to obtain the fault type of the air conditioning unit;
The plurality of sampling base sequences are base sequences which have been sampled before the current moment; the characteristic sequence prediction model is constructed based on a self-attention mechanism model; the feature sequence prediction model comprises: an inference encoder, a sampling encoder and a decoder; the sampling encoder is used for processing the plurality of sampling base sequences to obtain first result data; the inference encoder is used for processing the past known characteristic sequences to obtain second result data; the decoder is used for decoding the first result data, the second result data and the first feature sequence to obtain predicted values of all features in the second feature sequence; the base sequence comprises: component state information acquired in the running process of the air conditioning unit;
The step of inputting the plurality of sampling base sequences and the inference base sequence into a feature sequence prediction model to obtain predicted values of each feature in the second feature sequence, including:
Processing the plurality of sampling base sequences according to a first preset processing method, and inputting the processed sampling base sequences into the sampling encoder to obtain first result data;
processing the past known characteristic sequence according to a second preset processing method, and inputting the processed characteristic sequence into the inference encoder to obtain second result data;
processing the first characteristic sequence according to the second preset processing method to obtain third result data;
the first preset processing method is used for reducing the number dimension and the feature dimension of the sampling base sequence; the second preset processing method is used for reducing the feature dimension of the feature sequence; the sampling encoder and the inference encoder are comprised of a first number of encoder layers;
The step of inputting the plurality of sampling base sequences and the inference base sequence into a feature sequence prediction model to obtain predicted values of each feature in the second feature sequence, including:
Inputting the first result data and the second result data into the decoder as a key K and a value V after being spliced, and inputting the third result data into the decoder as a query Q and the key K and the value V at the same time, and predicting the values of all the features in the second feature sequence to obtain predicted values of all the features in the second feature sequence;
Wherein the decoder is comprised of a second number of decoder layers;
Before the plurality of sampling base sequences and the inference base sequences are input into a feature sequence prediction model to obtain the predicted values of the features in the second feature sequence, the method further comprises:
acquiring a plurality of sample sampling base sequences as training samples to be processed, and carrying out normalization processing on the training samples according to a maximum and minimum normalization method to obtain target training samples;
training a self-attention mechanism model by using the target training sample to obtain the characteristic sequence prediction model;
In the training process, the mean square error is used as a loss function to realize the reverse propagation and the updating of the model parameters.
2. The method according to claim 1, wherein the obtaining a plurality of sampling base sequences and an inference base sequence corresponding to a current time instant includes:
Acquiring standard base sequences acquired at a plurality of moments before the current moment to acquire a plurality of sampling base sequences;
Acquiring the past known feature sequences in the inference base sequences which are acquired before the current moment, and constructing the inference base sequences based on all known features in the standard base sequences and all unknown features in the standard base sequences after the current moment;
The standard base sequence comprises feature data corresponding to a plurality of operation stages, and the feature quantity of the feature data corresponding to each stage is a first feature quantity; each feature in the second sequence of features is filled with a zero value.
3. The method of claim 2, wherein the processing the plurality of sample base sequences according to the first preset processing method, and inputting the processed sample base sequences into the sample encoder to obtain first result data, includes:
Reducing the number dimension of the plurality of sampling base sequences to a preset number dimension through a full connection layer to obtain a first low-dimensional data;
And reducing the characteristic dimension of the first low-dimensional data to a preset characteristic dimension through a full connection and position coding layer to obtain second low-dimensional data, and adding position coding information to the second low-dimensional data to obtain the first result data.
4. A method according to claim 3, wherein said processing said past known signature sequence according to a second predetermined processing method is then input to said inference encoder to obtain second result data, comprising:
reducing the feature dimension of the past known feature sequence to a preset feature dimension through a full connection and position coding layer to obtain third low-dimensional data, and adding position coding information to the third low-dimensional data to obtain second result data;
and after the first feature sequence is processed according to the second preset processing method, obtaining third result data, wherein the third result data comprises:
And reducing the feature dimension of the first feature sequence to a preset feature dimension through a full connection and position coding layer to obtain fourth low-dimensional data, and adding position coding information to the fourth low-dimensional data to obtain third result data.
5. An air conditioning unit fault type diagnosis system based on big data, the system comprising:
The acquisition module is used for acquiring a plurality of sampling base sequences and reasoning base sequences corresponding to the current moment; the inference base sequence includes: a past known feature sequence, a future known first feature sequence, and a future unknown second feature sequence; the second characteristic sequence is a characteristic sequence to be inferred;
the prediction module is used for inputting the plurality of sampling base sequences and the reasoning base sequences into a feature sequence prediction model to obtain predicted values of all features in the second feature sequence; the predicted value is used for representing component state information of the air conditioning unit;
The fault type diagnosis module is used for diagnosing the air conditioning unit based on the predicted value of each feature in the second feature sequence to obtain the fault type of the air conditioning unit;
The plurality of sampling base sequences are base sequences which have been sampled before the current moment; the characteristic sequence prediction model is constructed based on a self-attention mechanism model; the feature sequence prediction model comprises: an inference encoder, a sampling encoder and a decoder; the sampling encoder is used for processing the plurality of sampling base sequences to obtain first result data; the inference encoder is used for processing the past known characteristic sequences to obtain second result data; the decoder is used for decoding the first result data, the second result data and the first feature sequence to obtain predicted values of all features in the second feature sequence; the base sequence comprises: component state information acquired in the running process of the air conditioning unit;
The step of inputting the plurality of sampling base sequences and the inference base sequence into a feature sequence prediction model to obtain predicted values of each feature in the second feature sequence, including:
Processing the plurality of sampling base sequences according to a first preset processing method, and inputting the processed sampling base sequences into the sampling encoder to obtain first result data;
processing the past known characteristic sequence according to a second preset processing method, and inputting the processed characteristic sequence into the inference encoder to obtain second result data;
processing the first characteristic sequence according to the second preset processing method to obtain third result data;
the first preset processing method is used for reducing the number dimension and the feature dimension of the sampling base sequence; the second preset processing method is used for reducing the feature dimension of the feature sequence; the sampling encoder and the inference encoder are comprised of a first number of encoder layers;
The step of inputting the plurality of sampling base sequences and the inference base sequence into a feature sequence prediction model to obtain predicted values of each feature in the second feature sequence, including:
Inputting the first result data and the second result data into the decoder as a key K and a value V after being spliced, and inputting the third result data into the decoder as a query Q and the key K and the value V at the same time, and predicting the values of all the features in the second feature sequence to obtain predicted values of all the features in the second feature sequence;
Wherein the decoder is comprised of a second number of decoder layers;
Before the plurality of sampling base sequences and the inference base sequences are input into a feature sequence prediction model to obtain the predicted values of the features in the second feature sequence, the method further comprises:
acquiring a plurality of sample sampling base sequences as training samples to be processed, and carrying out normalization processing on the training samples according to a maximum and minimum normalization method to obtain target training samples;
training a self-attention mechanism model by using the target training sample to obtain the characteristic sequence prediction model;
In the training process, the mean square error is used as a loss function to realize the reverse propagation and the updating of the model parameters.
6. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the big data based air conditioning unit fault type diagnosis method according to any of claims 1 to 4 when executing the program.
7. A computer-readable storage medium, on which a computer program is stored, which when executed by a processor implements the big data based air conditioning unit fault type diagnosis method according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311656595.1A CN117592571B (en) | 2023-12-05 | 2023-12-05 | Air conditioning unit fault type diagnosis method and system based on big data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311656595.1A CN117592571B (en) | 2023-12-05 | 2023-12-05 | Air conditioning unit fault type diagnosis method and system based on big data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117592571A CN117592571A (en) | 2024-02-23 |
| CN117592571B true CN117592571B (en) | 2024-05-17 |
Family
ID=89913167
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311656595.1A Active CN117592571B (en) | 2023-12-05 | 2023-12-05 | Air conditioning unit fault type diagnosis method and system based on big data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117592571B (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115018196A (en) * | 2022-06-30 | 2022-09-06 | 杭州电子科技大学 | Turnout switch machine fault prediction method based on adjustable input LSTM model |
| CN115186904A (en) * | 2022-07-13 | 2022-10-14 | 清华大学 | Industrial equipment fault prediction method and device based on Transformer |
| CN116008729A (en) * | 2023-01-03 | 2023-04-25 | 国网安徽省电力有限公司池州供电公司 | Power failure diagnosis early warning method, device, equipment and storage medium |
| CN116089885A (en) * | 2023-02-21 | 2023-05-09 | 南京信息工程大学 | Unknown fault type diagnosis method based on random depth forest |
| CN116189053A (en) * | 2022-10-09 | 2023-05-30 | 北京邮电大学 | Multi-mode-based untrimmed long video identification system and method |
| CN116186547A (en) * | 2023-04-27 | 2023-05-30 | 深圳市广汇源环境水务有限公司 | Method for rapidly identifying abnormal data of environmental water affair monitoring and sampling |
| CN116629361A (en) * | 2023-03-09 | 2023-08-22 | 南京大学 | Knowledge reasoning method based on ontology learning and attention mechanism |
| CN116720158A (en) * | 2023-04-26 | 2023-09-08 | 中国科学院合肥物质科学研究院 | Time sequence regression prediction method and system with uncertainty estimation |
| CN117009770A (en) * | 2023-08-30 | 2023-11-07 | 安徽大学 | Bearing fault diagnosis method based on SDP and visual transducer codes |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20220319157A1 (en) * | 2021-04-06 | 2022-10-06 | Nec Laboratories America, Inc. | Temporal augmentation for training video reasoning system |
| US20230252139A1 (en) * | 2022-02-10 | 2023-08-10 | Nec Laboratories America, Inc. | Efficient transformer for content-aware anomaly detection in event sequences |
-
2023
- 2023-12-05 CN CN202311656595.1A patent/CN117592571B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115018196A (en) * | 2022-06-30 | 2022-09-06 | 杭州电子科技大学 | Turnout switch machine fault prediction method based on adjustable input LSTM model |
| CN115186904A (en) * | 2022-07-13 | 2022-10-14 | 清华大学 | Industrial equipment fault prediction method and device based on Transformer |
| CN116189053A (en) * | 2022-10-09 | 2023-05-30 | 北京邮电大学 | Multi-mode-based untrimmed long video identification system and method |
| CN116008729A (en) * | 2023-01-03 | 2023-04-25 | 国网安徽省电力有限公司池州供电公司 | Power failure diagnosis early warning method, device, equipment and storage medium |
| CN116089885A (en) * | 2023-02-21 | 2023-05-09 | 南京信息工程大学 | Unknown fault type diagnosis method based on random depth forest |
| CN116629361A (en) * | 2023-03-09 | 2023-08-22 | 南京大学 | Knowledge reasoning method based on ontology learning and attention mechanism |
| CN116720158A (en) * | 2023-04-26 | 2023-09-08 | 中国科学院合肥物质科学研究院 | Time sequence regression prediction method and system with uncertainty estimation |
| CN116186547A (en) * | 2023-04-27 | 2023-05-30 | 深圳市广汇源环境水务有限公司 | Method for rapidly identifying abnormal data of environmental water affair monitoring and sampling |
| CN117009770A (en) * | 2023-08-30 | 2023-11-07 | 安徽大学 | Bearing fault diagnosis method based on SDP and visual transducer codes |
Non-Patent Citations (3)
| Title |
|---|
| PeiFeng Liu 等.The joint knowledge reasoning model based on knowledge representation learning for aviation assembly domain.《Science China Technological Sciences 》.2023,第143-156页. * |
| 张士刚 ; 刘颖 ; 宋立军 ; 胡政 ; 杨拥民 ; .基于模型的系统级故障诊断与预测推理技术.国防科技.2016,(02),第14-18页. * |
| 张泽生 ; 李宏光 ; 杨博 ; 张菁 ; .基于置信规则库推理的过程报警预测方法.控制工程.2019,(04),第155-162页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117592571A (en) | 2024-02-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110555479B (en) | Fault feature learning and classifying method based on 1DCNN and GRU fusion | |
| CN110610035B (en) | Rolling bearing residual life prediction method based on GRU neural network | |
| Zhou et al. | A novel grey prognostic model based on Markov process and grey incidence analysis for energy conversion equipment degradation | |
| CN108256173B (en) | Gas circuit fault diagnosis method and system for dynamic process of aircraft engine | |
| CN115791174B (en) | Rolling bearing abnormality diagnosis method, system, electronic equipment and storage medium | |
| Chen et al. | Hierarchical method for wind turbine prognosis using SCADA data | |
| CN111723925A (en) | Method, device, equipment and medium for fault diagnosis of on-road intelligent train air conditioning unit | |
| CN115455746B (en) | Nuclear power device operation monitoring data anomaly detection and correction integrated method | |
| CN115828466A (en) | Fan main shaft component fault prediction method based on wide kernel convolution | |
| Yan et al. | Deep learning technology for chiller faults diagnosis | |
| Zhu et al. | Res-HSA: Residual hybrid network with self-attention mechanism for RUL prediction of rotating machinery | |
| Huo et al. | A class-level matching unsupervised transfer learning network for rolling bearing fault diagnosis under various working conditions | |
| Sun et al. | Deep learning framework for gas turbine performance digital twin and degradation prognostics from airline operator perspective | |
| CN111766513B (en) | Capsule network-based variable-working-condition multi-fault diagnosis method for three-phase induction motor | |
| Wu et al. | Real-time monitoring and diagnosis scheme for IoT-enabled devices using multivariate SPC techniques | |
| CN117592571B (en) | Air conditioning unit fault type diagnosis method and system based on big data | |
| Khan et al. | A Review of machine learning techniques for wind turbine’s fault detection, diagnosis, and prognosis | |
| Souza et al. | Evaluation of data based normal behavior models for fault detection in wind turbines | |
| Song et al. | Anomaly detection of wind turbine generator based on temporal information | |
| Lin | Intelligent fault diagnosis of consumer electronics sensor in IoE via transformer | |
| CN117688496B (en) | Abnormality diagnosis method, system and equipment for satellite telemetry multidimensional time sequence data | |
| Wang et al. | Fault Diagnosis of Wind Turbine Generator with Stacked Noise Reduction Autoencoder Based on Group Normalization | |
| CN118035730B (en) | Equipment fault detection method, electronic equipment and storage medium | |
| CN117743875A (en) | Wind power converter fault early warning method, system and electronic equipment | |
| KR102212022B1 (en) | Method of automatically determining condition of hydro turbine in hydroelectric power plant and system for the same |
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 |