CN117216603A - Method and system for predicting faults of tubular falling film evaporator - Google Patents
Method and system for predicting faults of tubular falling film evaporator Download PDFInfo
- Publication number
- CN117216603A CN117216603A CN202311467588.7A CN202311467588A CN117216603A CN 117216603 A CN117216603 A CN 117216603A CN 202311467588 A CN202311467588 A CN 202311467588A CN 117216603 A CN117216603 A CN 117216603A
- Authority
- CN
- China
- Prior art keywords
- temperature
- value
- falling film
- film evaporator
- time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 239000011552 falling film Substances 0.000 title claims abstract description 84
- 238000000034 method Methods 0.000 title claims abstract description 52
- 238000005070 sampling Methods 0.000 claims abstract description 76
- 239000011159 matrix material Substances 0.000 claims abstract description 38
- 230000008859 change Effects 0.000 claims description 9
- 230000006870 function Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 6
- 238000013136 deep learning model Methods 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 abstract description 12
- 239000003507 refrigerant Substances 0.000 description 18
- 239000007788 liquid Substances 0.000 description 12
- 230000008569 process Effects 0.000 description 10
- 238000012545 processing Methods 0.000 description 7
- 238000001514 detection method Methods 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000009471 action Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000005484 gravity Effects 0.000 description 2
- 238000012546 transfer Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 238000004378 air conditioning Methods 0.000 description 1
- 238000009835 boiling Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000010612 desalination reaction Methods 0.000 description 1
- 238000005265 energy consumption Methods 0.000 description 1
- 239000010408 film Substances 0.000 description 1
- 238000012423 maintenance 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
- 230000000750 progressive effect Effects 0.000 description 1
- 238000005057 refrigeration Methods 0.000 description 1
- 239000013535 sea water Substances 0.000 description 1
- 239000007921 spray Substances 0.000 description 1
- 238000005507 spraying Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Landscapes
- Investigating Or Analyzing Materials Using Thermal Means (AREA)
Abstract
The invention relates to the technical field of fault prediction, in particular to a method and a system for predicting faults of a tubular falling film evaporator, which specifically comprise the following steps: temperature data are acquired through temperature sensors at the left side and the right side of the tubular falling film evaporator, and temperature divergence coefficients of all sampling moments are obtained according to temperature sequences at the left side and the right side of each sampling moment; calculating a cluster temperature highlighting sequence of a single temperature sensor in a time window by combining a divergence coefficient, constructing a left temperature window feature matrix and a right temperature window feature matrix by combining a temperature average value in the time cluster, obtaining a window temperature cross-correlation matrix by combining the two feature matrices, finally obtaining an evaporator fault index, realizing the prediction of the evaporator fault index by a neural network, and setting a fault threshold value to early warn the fault of the evaporator. Therefore, the fault prediction of the tubular falling film evaporator is realized, the problem of inaccurate fault prediction caused by other data interference is avoided, and the equipment fault can be processed in time.
Description
Technical Field
The invention relates to the technical field of fault prediction, in particular to a method and a system for predicting faults of a tube type falling film evaporator.
Background
The tube type falling film evaporator is a high-efficiency heat exchanger, has the advantages of high heat conduction coefficient, small heat transfer temperature difference, small refrigerant filling amount and the like, and is popularized and applied in the aspects of refrigeration air-conditioning, petrochemical industry, food processing, sea water desalination and the like. The tube type falling film evaporator is mainly characterized in that the refrigerant is sprayed to the surface of a heat exchange tube through a liquid distributor at the upper part of the evaporator, a layer of liquid refrigerant film is formed on the outer surface of the heat exchange tube, and then the refrigerant flows downwards step by step under the action of gravity to wrap all the heat exchange tubes, so that the net liquid level pressure of the refrigerant is reduced in the boiling heat exchange process, and the heat exchange efficiency is improved. Although the heat exchange efficiency of the tubular falling film evaporator is excellent, the liquid distributor is complex in structure and easy to cause equipment faults, so that uneven spraying of the liquid distributor is caused, the heat exchange efficiency is reduced, and the energy consumption is increased.
Because the shell of the tube type falling film evaporator is integrally formed, and the internal structure is complex, factors such as a liquid distributor, condensate, a heat exchange tube and the like are concentrated. On one hand, the interference of data obtained by the traditional detection mode is large, and the situation of false alarm or missing alarm exists, and on the other hand, the internal reaction situation can be obtained through the high-precision sensor, but the data obtaining cost is too large, so that the popularization of industry is not facilitated.
According to the invention, the temperature sensor is used for acquiring the surface temperature information of the shell of the tubular falling film evaporator, and the internal fault condition of the evaporator can be accurately obtained by analyzing the temperature and combining with neural network fitting, so that the fault detection and prediction of the tubular falling film evaporator are realized.
Disclosure of Invention
In order to solve the technical problems, the invention aims to provide a failure prediction method and a failure prediction system for a tubular falling film evaporator, and the adopted technical scheme is as follows:
in a first aspect, an embodiment of the present invention provides a method for predicting a failure of a tube falling film evaporator, including the steps of:
collecting temperature data of the tubular falling film evaporator, and taking a sequence formed by temperature values of all temperature sensors at the left side of the tubular falling film evaporator at each sampling moment as a left side temperature sequence at each sampling moment; acquiring a right temperature sequence of each sampling moment by combining the temperature value of the right temperature sensor;
taking the preset sampling time number as a time window, clustering all temperature values of the same temperature sensor through a DTC deep learning model for each time window to obtain each cluster, and marking the clusters as each time cluster; obtaining the temperature information disturbance degree of each sampling moment according to the left and right temperature sequences of each sampling moment; obtaining temperature divergence coefficients of all sampling moments according to the temperature information disorder degree of all the sampling moments; for each temperature sensor in each time cluster, obtaining the saliency of the temperature value of each sampling moment of the temperature sensor according to the change of the temperature value obtained by the temperature sensor; obtaining a temperature salient coefficient of a temperature sensor in the time cluster according to the temperature divergence coefficient and the saliency; for each temperature sensor, obtaining the inter-cluster temperature confusion coefficient of the temperature sensor in each time cluster according to the temperature highlighting coefficient; obtaining a left temperature window characteristic matrix according to the inter-cluster temperature confusion coefficients of each temperature sensor on the left side; combining the inter-cluster temperature confusion coefficients of the temperature sensors on the right side to obtain a right temperature window feature matrix; obtaining fault indexes of the tubular falling film evaporator according to the characteristic matrixes of the left temperature window and the right temperature window;
calculating all temperature averages in each time window, inputting the fault indexes and the temperature averages of the evaporators in all time windows into a long-period memory network, and outputting a fault index predicted value; and judging the fault condition of the tubular falling film evaporator according to the fault index predicted value.
Preferably, the obtaining the temperature information disturbance degree of each sampling time according to the left and right temperature sequences of each sampling time specifically includes:
for each sampling moment, respectively calculating the information entropy of all elements in the left and right temperature sequences, and respectively recording the information entropy as a first information entropy and a second information entropy; calculating the absolute value of the difference between the first information entropy and the second information entropy; obtaining a calculation result of an exponential function taking a natural constant as a base and taking the absolute value of the difference value as an index; and taking the calculation result as the temperature information disturbance degree at the sampling moment.
Preferably, the obtaining the temperature divergence coefficient of each sampling time according to the temperature information disorder degree of each sampling time specifically includes:
for each sampling moment, calculating the absolute value of the difference value of the temperature values of each temperature sensor on the left side and the temperature value of the corresponding temperature sensor on the right side; calculating the average value of all the absolute values of the differences; calculating the product of the temperature information disorder degree and the average value; and taking the product as a temperature divergence coefficient at the sampling moment.
Preferably, the obtaining the saliency of the temperature value at each sampling time of the temperature sensor according to the change of the temperature value obtained by the temperature sensor specifically includes:
calculating the average value and standard deviation of the residual temperature value after eliminating the temperature value at the ith sampling moment; calculating the difference between the temperature value at the ith sampling moment and the average value; calculating the absolute value of the ratio of the difference value to the standard deviation; and taking the absolute value of the ratio as the saliency of the temperature value at the ith sampling moment of the temperature sensor.
Preferably, the obtaining the temperature saliency coefficient of the temperature sensor in the time cluster according to the temperature bifurcation coefficient and the saliency specifically includes:
calculating the product of the salient coefficient and the temperature divergence coefficient of the temperature value at each sampling moment; calculating the average value of all the products; and taking the average value as a temperature highlighting coefficient of a temperature sensor in the time cluster.
Preferably, the obtaining the inter-cluster temperature confusion coefficient of the temperature sensor in each time cluster according to the temperature saliency coefficient specifically includes:
calculating the difference value of the temperature salient coefficients of the temperature sensors in the kth time cluster and other time clusters; acquiring a temperature time sequence of a temperature sensor in a kth time cluster; calculating a pearson correlation coefficient between the kth time cluster and the temperature time sequences of other time clusters; calculating the absolute value of the ratio of the difference value of the kth time cluster to the pearson correlation coefficient; calculating the average value of the absolute values of the ratios of all the time clusters; and taking the average value as an inter-cluster temperature confusion coefficient of the temperature sensor in the kth time cluster.
Preferably, the obtaining the left temperature window feature matrix according to the inter-cluster temperature confusion coefficient of each temperature sensor on the left side specifically includes:
for each temperature sensor on the left side, calculating the average value of all temperature values of the temperature sensors in each time cluster; calculating the product of the inter-cluster temperature confusion coefficient of the temperature sensor in the time cluster and the average value; taking the product as an element value;
and taking a matrix formed by all the element values as a left temperature window characteristic matrix.
Preferably, the obtaining the failure index of the tube type falling film evaporator according to the characteristic matrices of the left and right temperature windows specifically includes:
calculating the sum value of Frobenius norms of the characteristic matrixes of the left temperature window and the right temperature window; calculating a transposed matrix of the right temperature window feature matrix; calculating the Frobenius norm of the dot product of the left temperature window characteristic matrix and the transposed matrix; calculating the ratio of the sum to the Frobenius norm of the dot product; the ratio is taken as the failure index of the tube type falling film evaporator.
Preferably, the judging the fault condition of the tubular falling film evaporator according to the fault index predicted value specifically includes:
presetting a first threshold value and a second threshold value, wherein the first threshold value is smaller than the second threshold value; if the fault index predicted value is smaller than the first threshold value, the tubular falling film evaporator works normally; if the fault index predicted value is larger than or equal to the first threshold value and smaller than the second threshold value, the possibility of fault occurrence of the tubular falling film evaporator exists; if the fault index predicted value is larger than the second threshold value, the tube type falling film evaporator has faults.
In a second aspect, an embodiment of the present invention further provides a system for predicting a failure of a tube-type falling film evaporator, including a memory, a processor, and a computer program stored in the memory and running on the processor, where the processor implements the steps of any one of the methods described above when executing the computer program.
The embodiment of the invention has at least the following beneficial effects:
according to the invention, the temperature sensor is used for acquiring the surface temperature information of the shell of the tubular falling film evaporator, the temperature change condition of the shell of the tubular falling film evaporator is acquired by analyzing the temperature, the fault index of the evaporator is constructed, the equipment fault prediction is realized by combining a long-short-term memory network (LSTM) neural network, the data acquisition difficulty is reduced, the problem of inaccurate fault prediction caused by other data interference is avoided, and the equipment fault is treated in time;
temperature data are acquired through temperature sensors at the left side and the right side of the tubular falling film evaporator, and temperature bifurcation coefficients at all sampling moments are obtained according to temperature sequences at the left side and the right side of each sampling moment; calculating a temperature highlighting sequence of a single temperature sensor in a cluster in a time window by combining a divergence coefficient, constructing left and right temperature window feature matrixes by combining a temperature average value in the time cluster, obtaining a temperature cross-correlation matrix in the window by combining the two feature matrixes, finally obtaining an evaporator fault index, realizing the prediction of the evaporator fault index by a neural network, setting a fault threshold value, and carrying out early warning on the fault of the evaporator. Compared with the traditional mode, the method and the device have the advantages that equipment fault detection is realized only by placing the temperature sensor on the outer wall of the evaporator, the difficulty of data acquisition is reduced, in addition, equipment fault prediction is realized by combining the constructed equipment fault index through the neural network, and the timeliness of equipment processing is improved.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions and advantages of the prior art, the following description will briefly explain the drawings used in the embodiments or the description of the prior art, and it is obvious that the drawings in the following description are only some embodiments of the invention, and other drawings can be obtained according to the drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart illustrating steps of a method for predicting a failure of a tube type falling film evaporator according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a specific structure of a tube type falling film evaporator.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description refers to the specific implementation, structure, characteristics and effects of a method and a system for predicting the failure of a tube type falling film evaporator according to the invention by combining the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The invention provides a method and a system for predicting faults of a tubular falling film evaporator.
Referring to fig. 1, a flowchart of a method for predicting a failure of a tube type falling film evaporator according to an embodiment of the invention is shown, and the method includes the following steps:
the tube type falling film evaporator is widely used by virtue of the advantages of high heat exchange efficiency, small heat transfer temperature difference and small refrigerant filling amount, but the tube type falling film evaporator is complex in structure and easy to cause equipment failure, so that the tube type falling film evaporator needs to be subjected to failure detection and prediction. The specific structure schematic diagram of the tubular falling film evaporator is shown in fig. 2, the refrigerant enters the condensing pump from the high-pressure inlet, enters the interior of the evaporator from the refrigerant inlet, is uniformly sprayed on the heat exchange tube bundle through the refrigerant liquid distributor to realize heat exchange, flows into the liquid storage bag under the action of gravity, and finally flows back into the condensing pump to realize refrigerant circulation. The performance of the liquid distributor directly determines the liquid distribution uniformity of the refrigerant, so that the heat exchange performance of the falling film evaporator can be influenced, and the common sensor equipment is easy to corrode and cannot effectively monitor the fault condition of the falling film evaporator because the refrigerant is filled in the evaporator. Therefore, the invention realizes the prediction of the fault of the tubular falling film evaporator by placing the temperature sensor on the outer wall of the evaporator and analyzing the temperature. The method comprises the following specific steps:
and S001, acquiring temperature data of the left side and the right side of the tubular falling film evaporator through a temperature sensor.
In order to predict failure of the tube falling film evaporator, the temperature of the evaporator needs to be monitored. In this embodiment, 20 temperature sensors are placed on two sides of the evaporator shell at equal intervals, and 40 temperature values are obtained for a single sampling time and normalized. In order to be able to more accurately detect the evaporator temperature change, the present embodiment sets the sampling interval of the temperature sampler to 10ms. Because the temperature sensor on the shell of the tubular falling film evaporator can uninterruptedly acquire temperature information, when a period of time is acquired, the acquired temperature data amount is approaching infinity, and calculation and analysis are not utilized. Thus, it is necessary to divide the time window for temperatures approaching an infinite length, and the present embodiment sets a single time windowThe temperature variation and distribution within a single time window is analyzed. It should be noted that, the number of temperature sensors, the sampling interval and the size of the time window can be set by the practitioner, and the embodiment is not limited in particular.
The temperature of the evaporator is thus represented, within a single time window, byAnd->Respectively representing the left and right temperature sequences of the evaporator at the a-th sampling moment, wherein the left temperature sequence is +.>Right temperature sequence->，/>、/>The temperature values obtained by the m-th temperature sensors on the left and right sides of the evaporator at the a-th sampling time are respectively shown.
Step S002, calculating the temperature divergence coefficient of the evaporator at each sampling time by placing the temperatures obtained by the temperature sensors at the left side and the right side of the evaporator, and obtaining the failure index of the evaporator in the time window by combining the fluctuation condition of the temperatures in each time cluster in the time window.
In an ideal case, a uniform heat exchange takes place between the refrigerant inside the evaporator and the heat exchange jacket over a time window, so that the temperature values obtained are theoretically constant values, but in practice the heat exchange takes place in real time, with a certain fluctuation of the temperature obtained for the different sampling moments. Therefore, in this embodiment, a DTC deep learning model is adopted, and for each temperature sensor in a time window, temperature values at all sampling moments are adaptively clustered according to the fluctuation similarity of temperature, and because the DTC deep learning model is a known technology, the specific process is not described in detail in this embodiment, and therefore, the temperature values collected by each temperature sensor in a time window are divided into clusters, each cluster is determined as each time cluster, and the number of time clusters is determined as。
Through the mode, the temperature value of each temperature sensor is divided in time sequence, and subsequent processing analysis is convenient. Therefore, when the tube type falling film evaporator normally operates, the refrigerant is uniformly sprayed on the heat exchange tube bundle inside the evaporator through the liquid distributor, the refrigerant completely wraps the heat exchange tube bundle, and uniform heat exchange is realized, so that the temperature on the left side and the right side of the single sampling moment is a constant value theoretically.
However, certain differences may exist in the temperature of each position in the evaporator due to factors such as uneven heat distribution in the heat exchange tube bundle or refrigerant adhesion, so that in a single time window, the temperature divergence coefficient of the evaporator at each sampling moment is calculated according to the temperature sequences obtained by the temperature sensors at the left side and the right side of each sampling moment, and the expression is as follows:
in the method, in the process of the invention,for the temperature divergence coefficient of the evaporator at sample a, < >>For the temperature information disturbance of the evaporator at the a-th sampling time, < >>For the number of temperature sensors on the left or right side of the evaporator +.>，And->Temperature values of the m-th temperature sensor on the left and right side of the evaporator at the a-th sampling time, respectively,/->Andleft temperature sequence and right temperature sequence of a sampling time a are respectively, and +.>To calculate all of the sequencesThe information entropy of the elements is a function of the information entropy of the elements, which is a known technology, and the detailed calculation process is not repeated>Is an exponential function based on e.
When the tube type falling film evaporator works normally, the temperature of the corresponding positions of the temperature sensors at the left side and the right side of the single sampling moment is different, but the difference is small. When the evaporator works normally, the information entropy of the temperature sequences at the left side and the right side of the single sampling moment is relatively close, so that the obtained temperature information is disorderedThe value of (2) is smaller, in addition, the temperature difference between the temperature sensors at the left side and the right side of the same level is smaller, and finally the temperature divergence coefficient of the evaporator is obtained>The value of (2) is small. On the contrary, if the tubular falling film evaporator has unstable work or faults, the difference of the temperature sequences at the left side and the right side of the single sampling moment is larger, so that the final obtained +.>The value increases.
Temperature divergence coefficient of evaporator at single sampling timeReflecting the temperature uniformity of the whole evaporator at a single sampling moment, when the evaporator works normally, the heat is uniformly exchanged inside, so that the obtained temperature divergence coefficient is smaller. In addition, since the temperatures of the single-point positions of the evaporators have a correlation in terms of time transformation, it is necessary to analyze the temperature change conditions of each time cluster in a time window, wherein, taking any one of all the temperature sensors q on the left and right sides as an example, in one time window, the temperature sensor q is used as a reference>Temperature value representing the ith sampling instant of temperature sensor q in the kth time cluster。
Constructing a cluster temperature highlighting sequence of the temperature sensor q according to the temperature divergence coefficient of the evaporator at each sampling moment and the temperature value of the temperature sensor q changing along with time, wherein each element value in the sequence represents the temperature highlighting coefficient of the temperature sensor q in each time cluster, and calculating the temperature highlighting coefficient of the temperature sensor q in each time cluster by taking the kth time cluster as an example, wherein the expression is as follows:
in the method, in the process of the invention,for the temperature highlighting coefficient of the temperature sensor q in the kth time cluster, +.>For the number of sampling instants in the kth time cluster, is->For the temperature divergence coefficient of the evaporator in the kth time cluster at the ith sampling instant, for>For the salience of the temperature value of the temperature sensor q in the kth time cluster at the ith sampling instant, +.>For the temperature value of the temperature sensor q in the kth time cluster at the ith sampling instant, a.>And->Respectively, the average value of the residual temperature values after the temperature values at the ith sampling moment are removed from the kth time clusterAnd standard deviation.
The temperature fluctuation condition in the time cluster can be reflected, if the evaporator can work stably all the time, the temperature of a single temperature sensor in the time cluster is relatively constant, and the temperature should fluctuate in a small range near the mean value, so that the influence of the current temperature value on the mean value and the standard deviation in the window is eliminated, the difference is smaller, and the obtained temperature sensor is obtained>The value is small and at the same time the evaporator is operating steadily>Smaller, finally resulting in +.>The value is smaller; conversely, when the evaporator is not operating stably, the temperature fluctuations in the time cluster are large, eventually leading to +.>Increasing.
Window the timeThe number of inner time clusters is determined as +.>When all time clusters are traversed, the time window can be obtained>Intra-cluster temperature highlight sequence of intra-temperature sensor q>. This results in a time window->The temperature change in the respective time cluster, thus combining +.>The temperature highlight coefficients of each time cluster in the time cluster are used for obtaining the inter-cluster temperature confusion coefficients of the temperature sensor q in the corresponding time cluster, and the expression is as follows:
in the method, in the process of the invention,is the inter-cluster temperature confusion coefficient of the temperature sensor q in the kth time cluster, +.>For time window->Number of inner time clusters, +.>And->Temperature highlighting coefficient of temperature sensor q in jth and kth time cluster,/, respectively>And->Temperature time series of temperature sensor q in the jth and kth time cluster, respectively, +.>As a function of the calculated pearson correlation coefficient.
And calculating a corresponding inter-cluster temperature confusion coefficient for each time cluster in a single time window. When the liquid distributor in the evaporator can uniformly spray, the refrigerant and the heat exchange tube can exchange heat effectively, so that the temperature time sequence changes of all time clusters in the time window are consistent, and the correlation of the temperature time sequence is obtainedCoefficient of coefficientThe value is larger, in addition, the evaporator works stably, the temperature highlighting coefficient of the temperature sensor q in each time cluster is relatively close, and the finally obtained inter-cluster temperature confusion coefficient is->The value becomes smaller. On the contrary, the evaporator operation fluctuation is unstable, the difference of each time cluster is large, and the +.>The value increases.
In the time windowIn the method, a left temperature window characteristic matrix and a right temperature window characteristic matrix are constructed by combining inter-cluster temperature confusion coefficients of each temperature sensor in each time cluster, wherein the left temperature window characteristic matrix is taken as an example:
in the method, in the process of the invention,time window->Inside left temperature window feature matrix, +.>Is the inter-cluster temperature confusion coefficient of the mth temperature sensor at the left side of the evaporator in the kth time cluster, +.>For the mean value of the temperatures in the kth time cluster of the mth sensor on the left side of the evaporator,/>The number of temperature sensors on the left side of the evaporator is set according to the embodiment/>，/>Time window->Number of intra-time clusters.
In the temperature characteristic matrix,reflecting the difference in temperature between the kth time cluster and the remaining time clusters, and (2)>Reflecting the magnitude of the temperature in the time cluster. When the tube falling-film evaporator is operating normally, the temperature fluctuations can be kept at a small, approximately zero value by means of an efficient heat exchange, so that in the ideal case the temperature window characteristic matrix on the left side +.>The values of each element within are approaching zero and the values are not greatly different.
For each temperature sensor on the right side, a left temperature window characteristic matrix acquisition method is adopted to acquire a right temperature window characteristic matrix, and the method is usedRepresentation, whereby a time window is constructed in combination with the left and right temperature characteristic matrices>The internal temperature cross-correlation matrix is used for obtaining an evaporator fault index, and the expression is as follows:
in the method, in the process of the invention,for time window->Failure index of internal evaporator>And->Time window->Inside left and right temperature window feature matrix, < + >>For calculating the function of the matrix Frobenius norm (F-norm), +.>To calculate the function of the transposed matrix +.>Is a normalization function.
When the tubular falling film evaporator is in a normal working state, the elements in the characteristic matrix are all values approaching zero aiming at the temperature window, so that the F norms of the characteristic matrix are all values approaching zero. In addition, as the heat in the heat exchange tube bundle can be subjected to uniform heat exchange by the refrigerant, the temperature sensors on the left side and the right side are close to each other in a fluctuation manner in a time window, and have larger correlation, so that the F norm of a temperature cross-correlation matrix in the window is obtained, and finally the failure index of the evaporator is obtainedThe number of (2) is small. Conversely, when the evaporator fails, the fluctuation and difference of the temperatures in the window increases so that +.>The value increases.
And step S003, obtaining a fault index predicted value according to the obtained fault index of the evaporator and combining a neural network, and carrying out fault early warning on the tubular falling film evaporator through the fault index predicted value.
By the method, the fault index of the tubular falling film evaporator in each time window is obtainedThe larger the value, the greater the probability of failure of the tube falling film evaporator. Calculating the average value of all temperature data in each time window, and taking the time sequence consisting of the temperature average values of all time windows as a window temperature average value sequence +.>. In order to implement fault prediction of the evaporator, the embodiment adopts a long-short-term memory network (LSTM) to perform data prediction, where the long-short-term memory network is a known technology, and the specific process is not repeated. The inputs to the LSTM neural network are the evaporator failure index for all time windows +.>Window temperature mean sequence +.>Adam is adopted as an optimizer during training, mean square error is adopted as a loss function, and the output of the network is the predicted value +.>。
Presetting a first threshold valueSecond threshold->，/>It should be noted that->Is->The value of (2) can be set by the practitioner himself, this embodiment will +.>Is->The values of (2) are respectively set to +.>、/>. Evaporator failure index prediction value according to neural network prediction>When->When the tube type falling film evaporator is in the theoretical fluctuation range, the tube type falling film evaporator can work normally; when->When the tube type falling film evaporator is in fault, the possibility of occurrence of the tube type falling film evaporator is indicated, and relevant staff are prompted to pay attention to observation regularly; when->When the tube type falling film evaporator fails, temperature information is required to be submitted to equipment maintenance personnel, and the equipment is timely maintained and overhauled. Compared with the traditional mode, the embodiment realizes equipment fault detection only by placing the temperature sensor on the outer wall of the evaporator, reduces the difficulty of data acquisition, and in addition, predicts equipment faults by combining the constructed equipment fault index through a neural network, thereby improving the timeliness of equipment processing.
Based on the same inventive concept as the above method, the embodiment of the invention further provides a pipe-type falling film evaporator fault prediction system, which comprises a memory, a processor and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to realize the steps of any one of the above pipe-type falling film evaporator fault prediction methods.
In summary, the embodiment of the invention provides a method for predicting faults of a tubular falling film evaporator, which is characterized in that temperature information of the surface of a shell of the tubular falling film evaporator is obtained through a temperature sensor, the temperature change condition of the shell of the tubular falling film evaporator is obtained through temperature analysis, the fault index of the evaporator is constructed, equipment fault prediction is realized by combining a long-short-term memory network (LSTM) neural network, the difficulty of data acquisition is reduced, the problem of inaccurate fault prediction caused by other data interference is avoided, and the method is beneficial to timely processing equipment faults;
according to the embodiment, temperature data are obtained through temperature sensors at the left side and the right side of the tubular falling film evaporator, and temperature divergence coefficients at all sampling moments are obtained according to temperature sequences at the left side and the right side of each sampling moment; calculating a temperature highlighting sequence of a single temperature sensor in a cluster in a time window by combining a divergence coefficient, constructing left and right temperature window feature matrixes by combining a temperature average value in the time cluster, obtaining a temperature cross-correlation matrix in the window by combining the two feature matrixes, finally obtaining an evaporator fault index, realizing the prediction of the evaporator fault index by a neural network, setting a fault threshold value, and carrying out early warning on the fault of the evaporator. Compared with the traditional mode, the method and the device have the advantages that equipment fault detection is achieved only by placing the temperature sensor on the outer wall of the evaporator, the difficulty of data acquisition is reduced, in addition, equipment fault prediction is achieved through the neural network by combining the constructed equipment fault index, and the timeliness of equipment processing is improved.
It should be noted that: the sequence of the embodiments of the present invention is only for description, and does not represent the advantages and disadvantages of the embodiments. And the foregoing description has been directed to specific embodiments of this specification. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments.
The foregoing description of the preferred embodiments of the present invention is not intended to be limiting, but rather, any modifications, equivalents, improvements, etc. that fall within the principles of the present invention are intended to be included within the scope of the present invention.
Claims (10)
1. The fault prediction method for the tubular falling film evaporator is characterized by comprising the following steps of:
collecting temperature data of the tubular falling film evaporator, and taking a sequence formed by temperature values of all temperature sensors at the left side of the tubular falling film evaporator at each sampling moment as a left side temperature sequence at each sampling moment; acquiring a right temperature sequence of each sampling moment by combining the temperature value of the right temperature sensor;
taking the preset sampling time number as a time window, clustering all temperature values of the same temperature sensor through a DTC deep learning model for each time window to obtain each cluster, and marking the clusters as each time cluster; obtaining the temperature information disturbance degree of each sampling moment according to the left and right temperature sequences of each sampling moment; obtaining temperature divergence coefficients of all sampling moments according to the temperature information disorder degree of all the sampling moments; for each temperature sensor in each time cluster, obtaining the saliency of the temperature value of each sampling moment of the temperature sensor according to the change of the temperature value obtained by the temperature sensor; obtaining a temperature salient coefficient of a temperature sensor in the time cluster according to the temperature divergence coefficient and the saliency; for each temperature sensor, obtaining the inter-cluster temperature confusion coefficient of the temperature sensor in each time cluster according to the temperature highlighting coefficient; obtaining a left temperature window characteristic matrix according to the inter-cluster temperature confusion coefficients of each temperature sensor on the left side; combining the inter-cluster temperature confusion coefficients of the temperature sensors on the right side to obtain a right temperature window feature matrix; obtaining fault indexes of the tubular falling film evaporator according to the characteristic matrixes of the left temperature window and the right temperature window;
calculating all temperature averages in each time window, inputting the fault indexes and the temperature averages of the evaporators in all time windows into a long-period memory network, and outputting a fault index predicted value; and judging the fault condition of the tubular falling film evaporator according to the fault index predicted value.
2. The method for predicting faults of a tube type falling film evaporator according to claim 1, wherein the method for predicting faults of the tube type falling film evaporator is characterized by obtaining the temperature information disorder degree of each sampling moment according to the left and right temperature sequences of each sampling moment, and specifically comprises the following steps:
for each sampling moment, respectively calculating the information entropy of all elements in the left and right temperature sequences, and respectively recording the information entropy as a first information entropy and a second information entropy; calculating the absolute value of the difference between the first information entropy and the second information entropy; obtaining a calculation result of an exponential function taking a natural constant as a base and taking the absolute value of the difference value as an index; and taking the calculation result as the temperature information disturbance degree at the sampling moment.
3. The method for predicting faults of a tube type falling film evaporator according to claim 1, wherein the method for obtaining the temperature divergence coefficient of each sampling moment according to the temperature information disorder degree of each sampling moment comprises the following steps:
for each sampling moment, calculating the absolute value of the difference value of the temperature values of each temperature sensor on the left side and the temperature value of the corresponding temperature sensor on the right side; calculating the average value of all the absolute values of the differences; calculating the product of the temperature information disorder degree and the average value; and taking the product as a temperature divergence coefficient at the sampling moment.
4. The method for predicting the failure of the tube type falling film evaporator according to claim 1, wherein the obtaining the saliency of the temperature value of each sampling time of the temperature sensor according to the change of the temperature value obtained by the temperature sensor specifically comprises the following steps:
calculating the average value and standard deviation of the residual temperature value after eliminating the temperature value at the ith sampling moment; calculating the difference between the temperature value at the ith sampling moment and the average value; calculating the absolute value of the ratio of the difference value to the standard deviation; and taking the absolute value of the ratio as the saliency of the temperature value at the ith sampling moment of the temperature sensor.
5. The method for predicting faults of a tube type falling film evaporator according to claim 1, wherein the obtaining the temperature saliency coefficient of the temperature sensor in the time cluster according to the temperature bifurcation coefficient and the saliency comprises the following steps:
calculating the product of the salient coefficient and the temperature divergence coefficient of the temperature value at each sampling moment; calculating the average value of all the products; and taking the average value as a temperature highlighting coefficient of a temperature sensor in the time cluster.
6. The method for predicting failure of a tube type falling film evaporator according to claim 1, wherein the obtaining the inter-cluster temperature confusion coefficient of the temperature sensor in each time cluster according to the temperature saliency coefficient comprises the following steps:
calculating the difference value of the temperature salient coefficients of the temperature sensors in the kth time cluster and other time clusters; acquiring a temperature time sequence of a temperature sensor in a kth time cluster; calculating a pearson correlation coefficient between the kth time cluster and the temperature time sequences of other time clusters; calculating the absolute value of the ratio of the difference value of the kth time cluster to the pearson correlation coefficient; calculating the average value of the absolute values of the ratios of all the time clusters; and taking the average value as an inter-cluster temperature confusion coefficient of the temperature sensor in the kth time cluster.
7. The method for predicting faults of a tube type falling film evaporator according to claim 1, wherein the method for predicting faults of the tube type falling film evaporator is characterized by obtaining a left temperature window characteristic matrix according to inter-cluster temperature confusion coefficients of each temperature sensor on the left side, and specifically comprises the following steps:
for each temperature sensor on the left side, calculating the average value of all temperature values of the temperature sensors in each time cluster; calculating the product of the inter-cluster temperature confusion coefficient of the temperature sensor in the time cluster and the average value; taking the product as an element value;
and taking a matrix formed by all the element values as a left temperature window characteristic matrix.
8. The method for predicting the failure of the tube type falling film evaporator according to claim 1, wherein the method for obtaining the failure index of the tube type falling film evaporator according to the characteristic matrices of the left temperature window and the right temperature window comprises the following steps:
calculating the sum value of Frobenius norms of the characteristic matrixes of the left temperature window and the right temperature window; calculating a transposed matrix of the right temperature window feature matrix; calculating the Frobenius norm of the dot product of the left temperature window characteristic matrix and the transposed matrix; calculating the ratio of the sum to the Frobenius norm of the dot product; the ratio is taken as the failure index of the tube type falling film evaporator.
9. The method for predicting the failure of the tube type falling film evaporator according to claim 1, wherein the judging the failure condition of the tube type falling film evaporator according to the failure index predicted value specifically comprises the following steps:
presetting a first threshold value and a second threshold value, wherein the first threshold value is smaller than the second threshold value; if the fault index predicted value is smaller than the first threshold value, the tubular falling film evaporator works normally; if the fault index predicted value is larger than or equal to the first threshold value and smaller than the second threshold value, the possibility of fault occurrence of the tubular falling film evaporator exists; if the fault index predicted value is larger than the second threshold value, the tube type falling film evaporator has faults.
10. A tube falling film evaporator failure prediction system comprising a memory, a processor and a computer program stored in the memory and running on the processor, characterized in that the processor implements the steps of the method according to any one of claims 1-9 when the computer program is executed by the processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311467588.7A CN117216603B (en) | 2023-11-07 | 2023-11-07 | Method and system for predicting faults of tubular falling film evaporator |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311467588.7A CN117216603B (en) | 2023-11-07 | 2023-11-07 | Method and system for predicting faults of tubular falling film evaporator |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117216603A true CN117216603A (en) | 2023-12-12 |
| CN117216603B CN117216603B (en) | 2024-02-06 |
Family
ID=89044720
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311467588.7A Active CN117216603B (en) | 2023-11-07 | 2023-11-07 | Method and system for predicting faults of tubular falling film evaporator |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117216603B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117547843A (en) * | 2024-01-12 | 2024-02-13 | 河北绿鸿环保科技有限公司 | Comprehensive utilization method and system for strong brine and sintering machine head ash |
| CN117892095A (en) * | 2024-03-14 | 2024-04-16 | 山东泰开电力电子有限公司 | Intelligent detection method for faults of heat dissipation system for energy storage system |
| CN117704880B (en) * | 2023-12-13 | 2024-05-07 | 中建-大成建筑有限责任公司 | Energy-saving temperature monitoring method and system for evaporator |
| CN118090201A (en) * | 2024-04-26 | 2024-05-28 | 山东艾克索仑电气有限公司 | Mining permanent magnet synchronous motor-driven roller fault diagnosis method |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112731159A (en) * | 2020-12-23 | 2021-04-30 | 江苏省电力试验研究院有限公司 | Method for pre-judging and positioning battery fault of battery compartment of energy storage power station |
| CN112763929A (en) * | 2020-12-31 | 2021-05-07 | 华东理工大学 | Method and device for predicting health of battery monomer of energy storage power station system |
| CN113049142A (en) * | 2019-12-27 | 2021-06-29 | 华能如东八仙角海上风力发电有限责任公司 | Temperature sensor alarm method, device, equipment and storage medium |
| CN113312851A (en) * | 2021-06-16 | 2021-08-27 | 华电山东新能源有限公司 | Early warning method for temperature abnormity of main bearing of wind driven generator |
| CN115983452A (en) * | 2022-12-19 | 2023-04-18 | 河南科技大学 | Bearing fault prediction method based on LSTM |
| CN116879763A (en) * | 2023-09-07 | 2023-10-13 | 上海融和元储能源有限公司 | Battery fault early warning method based on Kalman filtering algorithm |
-
2023
- 2023-11-07 CN CN202311467588.7A patent/CN117216603B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113049142A (en) * | 2019-12-27 | 2021-06-29 | 华能如东八仙角海上风力发电有限责任公司 | Temperature sensor alarm method, device, equipment and storage medium |
| CN112731159A (en) * | 2020-12-23 | 2021-04-30 | 江苏省电力试验研究院有限公司 | Method for pre-judging and positioning battery fault of battery compartment of energy storage power station |
| CN112763929A (en) * | 2020-12-31 | 2021-05-07 | 华东理工大学 | Method and device for predicting health of battery monomer of energy storage power station system |
| CN113312851A (en) * | 2021-06-16 | 2021-08-27 | 华电山东新能源有限公司 | Early warning method for temperature abnormity of main bearing of wind driven generator |
| CN115983452A (en) * | 2022-12-19 | 2023-04-18 | 河南科技大学 | Bearing fault prediction method based on LSTM |
| CN116879763A (en) * | 2023-09-07 | 2023-10-13 | 上海融和元储能源有限公司 | Battery fault early warning method based on Kalman filtering algorithm |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117704880B (en) * | 2023-12-13 | 2024-05-07 | 中建-大成建筑有限责任公司 | Energy-saving temperature monitoring method and system for evaporator |
| CN117547843A (en) * | 2024-01-12 | 2024-02-13 | 河北绿鸿环保科技有限公司 | Comprehensive utilization method and system for strong brine and sintering machine head ash |
| CN117547843B (en) * | 2024-01-12 | 2024-03-15 | 河北绿鸿环保科技有限公司 | Comprehensive utilization method and system for strong brine and sintering machine head ash |
| CN117892095A (en) * | 2024-03-14 | 2024-04-16 | 山东泰开电力电子有限公司 | Intelligent detection method for faults of heat dissipation system for energy storage system |
| CN117892095B (en) * | 2024-03-14 | 2024-05-28 | 山东泰开电力电子有限公司 | Intelligent detection method for faults of heat dissipation system for energy storage system |
| CN118090201A (en) * | 2024-04-26 | 2024-05-28 | 山东艾克索仑电气有限公司 | Mining permanent magnet synchronous motor-driven roller fault diagnosis method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117216603B (en) | 2024-02-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117216603B (en) | Method and system for predicting faults of tubular falling film evaporator | |
| CN110441065B (en) | Gas turbine on-line detection method and device based on LSTM | |
| CN108803520B (en) | Dynamic process monitoring method based on variable nonlinear autocorrelation rejection | |
| Li et al. | A novel data-temporal attention network based strategy for fault diagnosis of chiller sensors | |
| CN109146246B (en) | Fault detection method based on automatic encoder and Bayesian network | |
| CN104914723B (en) | Industrial process soft-measuring modeling method based on coorinated training partial least square model | |
| CN111275288A (en) | XGboost-based multi-dimensional data anomaly detection method and device | |
| CN108469805B (en) | Distributed dynamic process monitoring method based on dynamic optimal selection | |
| CN116418120B (en) | Intelligent early warning method applied to water-cooled power supply | |
| CN109669415A (en) | A kind of dynamic process monitoring method based on the analysis of structuring canonical variable | |
| CN111352977A (en) | Time sequence data monitoring method based on self-attention bidirectional long-short term memory network | |
| CN106971027B (en) | Water chilling unit fault feature selection method based on DR-BN model | |
| CN104199441A (en) | Blast furnace multiple working condition fault separation method and system based on sparse contribution plot | |
| CN114580288A (en) | Intelligent fault diagnosis method based on DSECJAN | |
| CN108181893B (en) | PCA-KDR-based fault detection method | |
| CN116073436A (en) | Capacity optimization control method for photovoltaic new energy power system | |
| CN115204272A (en) | Industrial system fault diagnosis method and equipment based on multi-sampling rate data | |
| CN111913415B (en) | Continuous stirring reaction kettle operation state monitoring method based on time sequence data analysis | |
| CN116720150A (en) | Mechanical refrigeration system fault diagnosis method and system | |
| CN108536943B (en) | Fault monitoring method based on multi-production-unit variable cross-correlation decoupling strategy | |
| CN111914886B (en) | Nonlinear chemical process monitoring method based on online brief kernel learning | |
| CN114861759A (en) | Distributed training method of linear dynamic system model | |
| CN111914471A (en) | Rectification tower fault detection method based on rapid nuclear independent component analysis | |
| CN118035929B (en) | Power plant safety monitoring and early warning method and system based on thermodynamics and data mining | |
| Chang et al. | Soft sensor of the key effluent index in the municipal wastewater treatment process based on transformer |
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 |