Abnormity detection and quantitative diagnosis method and system for heat exchanger under multiple unstable working conditions, medium and equipment

Abstract

The invention relates to the field of heat exchanger anomaly detection and fault diagnosis, and discloses a heat exchanger anomaly detection and quantitative diagnosis method, system, medium and equipment under multiple unstable working conditions, and the method comprises the steps: collecting data of key parts of a heat exchanger, and preprocessing the data, the data including online monitoring data and offline data; according to the preprocessed data, fluctuation working condition sensitive indexes are constructed, the fluctuation working condition sensitive indexes comprise leakage fault indexes and scaling fault indexes, and follow-up characteristic analysis indexes of the online monitoring data are obtained through the fluctuation working condition sensitive indexes so as to construct an unsteady-state multi-working-condition anomaly detection model; leakage fault quantitative diagnosis and scaling fault quantitative diagnosis are carried out according to the abnormal detection model under the unsteady-state multiple working conditions. Abnormality detection of the heat exchanger under multiple unstable working conditions can be achieved, and the accuracy of abnormality detection is improved.

Description

Technical Field

[0001] The present invention relates to the field of abnormality detection and fault diagnosis of heat exchangers, and in particular to a method, system, medium and equipment for abnormality detection and quantitative diagnosis of heat exchangers under non-steady-state multiple working conditions. Background Art

[0002] As an efficient, energy-saving and compact heat exchange equipment, heat exchangers are currently widely used in industry. With the development of instrumentation and control technology, online monitoring technology has been widely used in the operation and maintenance of heat exchangers. However, there are few studies on intelligent operation and maintenance methods such as state perception detection and diagnosis and prediction of heat exchangers. The online failure identification, diagnosis and prediction of heat exchangers face the lack of abnormal identification, failure quantification and prediction methods under complex working conditions and complex failure situations.

[0003] The current mainstream research on heat exchanger anomaly detection technology is mostly based on laboratory environments, while heat exchangers actually put into operation have more complex operating conditions. Its complex parameter fluctuations and operating conditions will greatly affect the model accuracy of statistical anomaly detection methods. In addition, most heat exchangers actually put into operation will have the problem of lack of fault data, which greatly reduces the application scope of machine learning models and lacks anomaly detection methods under complex working conditions. On the other hand, heat exchanger fault diagnosis technology has made certain progress in the direction of mechanism, signal processing and data-driven, and most of them can achieve accurate fault identification and diagnosis. However, the current heat exchanger fault diagnosis methods are mostly focused on the identification and location of fault modes, while quantitative analysis is often required for fault modes such as leakage and scaling to guide operation and maintenance optimization decisions, and there is a lack of methods for quantitative fault diagnosis. Summary of the invention

[0004] In view of the above problems, the purpose of the present invention is to provide a method, system, medium and equipment for detecting and quantitatively diagnosing abnormalities in heat exchangers under non-steady-state multiple operating conditions, which can realize abnormality detection of heat exchangers under non-steady-state multiple operating conditions and improve the accuracy of abnormality detection.

[0005] To achieve the above-mentioned purpose, the present invention adopts the following technical scheme: a method for abnormality detection and quantitative diagnosis of heat exchangers under non-steady-state multi-operating conditions, which comprises: collecting data of key components of the heat exchanger and preprocessing the data, wherein the data of the key components include online monitoring data and offline data; constructing a fluctuating condition sensitive index based on the preprocessed data, the fluctuating condition sensitive index includes a leakage fault index and a scaling fault index, obtaining a follow-up characteristic analysis index of the online monitoring data from the fluctuating condition sensitive index to construct an abnormality detection model under non-steady-state multi-operating conditions; performing quantitative diagnosis of leakage faults and quantitative diagnosis of scaling faults according to the abnormality detection model under non-steady-state multi-operating conditions.

[0006] Further, the online monitoring parameters include temperature, pressure and volume flow;

[0007] Offline data includes structural parameters and physical parameters. The structural parameters include the equivalent diameter of the cold / hot side flow channel, the flow channel cross-sectional area, the flow channel length, the number of processes, the total heat transfer area, and the pipe diameter ratio; the physical parameters include the absolute viscosity, density and constant pressure specific heat of the cold / hot side medium.

[0008] Furthermore, a fluctuation condition sensitive index is constructed based on the preprocessed data, and a follow-up characteristic analysis index of the online monitoring data is obtained from the fluctuation condition sensitive index, including:

[0009] Assume that the heat exchanger inlet parameter is C2 and the outlet parameter is C1; the heat exchanger outlet parameter C2 changes with the change of the inlet parameter C1, and the specific change is determined by the hidden state of the heat exchanger;

[0010] When the hidden state changes, the heat exchanger deviates from the original working state and an abnormality occurs; the following characteristics of the parameters reflect the hidden state of the heat exchanger. By analyzing the following characteristics of the monitoring parameters, the hidden state of the heat exchanger can be inferred, thereby achieving abnormality detection.

[0011] Furthermore, an abnormality detection model under non-steady-state multi-operating conditions is constructed, including:

[0012] A health memory matrix of multiple operating conditions is constructed, and the spatial correlation characteristics between matrices are used to simplify the health memory matrix to achieve multi-operating condition and non-stationary online anomaly detection of heat exchangers.

[0013] Furthermore, quantitative diagnosis of leakage faults includes: starting from the perspective of heat transfer and energy conservation, constructing a quantitative diagnosis model of medium internal leakage and flange sealing surface leakage based on heat balance to perform quantitative diagnosis of medium leakage.

[0014] Furthermore, the quantitative diagnosis model for medium internal leakage and flange sealing surface leakage is:

[0015]

[0016] Where x is the leakage; V h is the thermal fluid flow rate; V c is the cold fluid flow rate; T hi is the tube inlet temperature; T ho is the tube outlet temperature, T ci is the shell inlet temperature, T co is the shell outlet temperature.

[0017] Furthermore, the quantitative diagnosis of scaling faults includes: constructing a scaling amount estimation method based on a benchmark heat transfer model, introducing a nonlinear regression model to construct a quantitative method for scaling thickness, and constructing a quantitative diagnostic model for scaling amount in combination with the response data of the actual system, which is:

[0018] δ f =R f ·λ f

[0019] In the formula, δ f is the thickness of the dirt layer; R f is the thermal resistance of the dirt layer; w is the thermal conductivity of the dirt layer.

[0020] A heat exchanger abnormality detection and quantitative diagnosis system under non-steady-state multiple working conditions, comprising: a data processing module, which collects data of key components of the heat exchanger and pre-processes the data, wherein the data includes online monitoring data and offline data; a following characteristic module, which constructs a fluctuation condition sensitive index based on the pre-processed data, the fluctuation condition sensitive index includes a leakage fault index and a scaling fault index, and obtains the following characteristic analysis index of the online monitoring data from the fluctuation condition sensitive index to construct an abnormality detection model under non-steady-state multiple working conditions; a diagnosis module, which performs leakage fault quantitative diagnosis and scaling fault quantitative diagnosis according to the abnormality detection model under non-steady-state multiple working conditions.

[0021] A computer-readable storage medium storing one or more programs, wherein the one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any of the above methods.

[0022] A computing device comprises: one or more processors, a memory and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any of the above methods.

[0023] Due to the adoption of the above technical solution, the present invention has the following advantages compared with the existing abnormality detection and fault diagnosis methods:

[0024] 1. In view of the problem that the anomaly detection model under the influence of variable working conditions is difficult to accurately distinguish between anomalies and normal conditions, the present invention introduces the idea of ​​cointegration analysis, and transforms the non-steady-state multi-condition anomaly detection problem into a steady-state multi-condition anomaly detection problem by eliminating parameter fluctuations, and constructs a multi-condition state monitoring and anomaly detection model. The present invention can realize the anomaly detection of heat exchangers under non-steady-state multi-conditions and improve the accuracy of anomaly detection.

[0025] 2. In order to solve the problem that the failure modes of equipment cooling water heat exchangers are diverse and difficult to predict, the present invention constructs a mechanism-driven fault identification and location inference engine, and introduces the idea of ​​virtual perception. For leakage failures, a heat conservation-based inference engine is constructed.

[0026] For scaling failure, a scaling estimation method based on a benchmark heat transfer model is constructed, and a nonlinear regression model is introduced to construct a quantitative method for scaling thickness. The present invention can achieve quantitative analysis of the degree of failure and reduce the error of fault diagnosis.

[0027] 3. The present invention sets up a state monitoring, diagnosis and prediction system for equipment cooling water heat exchangers. The system function requirements are analyzed from the perspectives of system functions and business functions, and functional modules such as equipment parameter management, state monitoring, fault analysis and management are designed and developed, realizing equipment characteristic parameter configuration, data storage, data and analysis result visualization, and report management. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a framework diagram of a multivariate state abnormality detection method based on following characteristics, taking a heat exchanger as an example in an embodiment of the present invention;

[0029] Figure 2 This is a framework diagram of a quantitative diagnosis method for a heat exchanger, taking a heat exchanger as an example in an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In view of the problem that the abnormality detection model is difficult to accurately distinguish abnormalities from normal conditions under the influence of variable working conditions, and the problem that the failure modes of equipment cooling water heat exchangers are diverse and difficult to measure, the present invention provides a method, system, medium and equipment for abnormality detection and quantitative diagnosis of heat exchangers under non-steady-state multi-working conditions, and a diagnosis model for industrial equipment groups based on cluster analysis. First, the idea of ​​cointegration analysis is introduced, and a multivariate state estimation abnormality detection method based on following characteristics is adopted. By studying the correlation between working condition fluctuations and state changes during the normal operation of the heat exchanger, the steady-state state quantity time series representing the equipment operating state in the non-steady-state working condition sequence is extracted, and a health memory matrix of multiple working conditions is constructed. The health memory matrix is ​​simplified by using the spatial correlation characteristics between matrices, and the multi-condition, non-stationary online abnormality detection of heat exchangers is realized. Finally, the qualitative fault identification and location technology and the quantitative fault degree analysis technology are combined to establish a heat exchanger fault diagnosis model based on qualitative and quantitative combination. The classification and location of faults are realized by establishing a qualitative fault diagnosis expert system. Then, through the mechanism derivation of the failure mode, a failure degree quantification method is constructed to realize accurate diagnosis of complex failures.

[0031] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field belong to the scope of protection of the present invention.

[0032] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0033] In one embodiment of the present invention, a method for detecting and diagnosing abnormalities in a heat exchanger under non-steady-state conditions is provided. Figure 1 , Figure 2 As shown, the method comprises the following steps:

[0034] 1) Collect data of key components of the heat exchanger, including online monitoring data and offline data. The online monitoring data includes process monitoring parameters such as temperature, pressure, and flow rate. The offline data includes heat exchanger equipment structure parameters and medium physical property parameters, and pre-process the data.

[0035] The data preprocessing includes:

[0036] 1.1) Online monitoring data collection and processing.

[0037] Online monitoring parameters, including temperature, pressure, volume flow, etc., are collected in real time through sensors. In order to make the model more accurate, online data needs to be preprocessed to eliminate the impact of the environment. The exponential moving average model is used to smooth the data:

[0038]

[0039] In the formula, EMA N (x n ) is the original data x n EMA value of x n-k is the raw data value at time nk; x n is the raw data value at time n, k is less than n; N is the number of periods of EMA.

[0040] 1.2) Offline data collection.

[0041] Offline data is mainly used to construct high-level monitoring indicators such as heat transfer coefficient and heat transfer efficiency. Therefore, it is necessary to focus on offline parameters related to high-level monitoring indicators, among which structural parameters include equivalent diameter of cold / hot side flow channel, flow channel cross-sectional area, flow channel length, number of processes, total heat transfer area, etc. For shell and tube heat exchangers, it is also necessary to know the tube diameter ratio, etc.; physical parameters mainly include absolute viscosity, density, constant pressure specific heat, etc. of cold / hot side media.

[0042] 2) Based on the preprocessed data, sensitive indicators of fluctuating conditions are constructed. The sensitive indicators of fluctuating conditions include leakage fault indicators and scaling fault indicators. The following characteristic analysis indicators of online monitoring data are obtained from the sensitive indicators of fluctuating conditions to construct an abnormality detection model under non-steady-state multi-conditions.

[0043] Among them, the fluctuation condition sensitive index is constructed according to the preprocessed data, and the following characteristic analysis index of the online monitoring data is obtained from the fluctuation condition sensitive index, which is specifically:

[0044] Assume that the heat exchanger inlet parameter is C2 and the outlet parameter is C1. The heat exchanger outlet parameter C2 changes with the change of the inlet parameter C1, and the specific change is determined by the hidden state of the device. When the hidden state changes, it means that the device has deviated from the original working state, that is, an abnormality has occurred. The following characteristics of the parameters reflect the hidden state of the device. By analyzing the following characteristics of the monitoring parameters, the hidden state of the device can be inferred to achieve abnormality detection. Among them, the device refers to the heat exchanger.

[0045] In this embodiment, the fluctuation condition sensitive index includes a leakage fault index and a scaling fault index. Specifically:

[0046] 2.1) Construction of leakage fault index. Leakage fault means that the fluid on one side flows through the leakage port to the flow channel on the other side or the outside of the equipment. Without considering the compressibility of the fluid, according to the continuity equation, before the leakage, we have:

[0047] G co =G ci

[0048] G ho =G hi

[0049] In the formula, G co is the cold flow outlet flow rate; G ci is the cold flow inlet flow rate; G ho is the heat flow outlet flow rate; G hi is the heat inlet flow rate.

[0050] When internal leakage occurs (assuming hot side cold test leakage):

[0051] G co -x=Gci

[0052] G ho +x=G hi

[0053] Where x is the leakage amount.

[0054] When external leakage occurs (assuming the hot side leaks outward):

[0055] G ho +x=G hi

[0056] After leakage occurs on the leakage side, the total flow area changes. According to the power calculation formula:

[0057] P=FV

[0058] Among them, P is the flow power consumption rate, the unit is W; F is the medium thrust, the unit is N; V is the flow velocity, the unit is m / s.

[0059] In summary, construct the following characteristic indicator:

[0060]

[0061] Among them, x c,g is the cold flow inlet and outlet flow difference; x h,g is the heat flow inlet and outlet flow difference; x h,p is the heat flow inlet and outlet pressure difference; x c,p P is the pressure difference between the inlet and outlet of the cold flow; hi is the heat flow inlet pressure; P ho is the hot flow outlet pressure; P ci is the cold flow inlet pressure; P co is the cold flow outlet pressure.

[0062] 2.2) Construction of scaling fault indicators.

[0063] Assuming that there is a local convection heat transfer flow field inside the heat exchanger, according to the convection heat transfer Newton cooling formula:

[0064] Q=h·ΔT

[0065] Where Q is the heat flux per unit area, in kW / m 2 ; h is the convective heat transfer coefficient, unit is kW / m 2 ·℃; ΔT is the temperature difference between the solid surface and the fluid, in ℃. Temporarily assume that the temperature of the wall is fixed, according to the law of heat conservation:

[0066]

[0067] Where, T pi and Tpo Respectively represent the temperature of the inflow and outflow sections of the flow field; T w is the wall temperature; C p is the specific heat capacity at constant pressure; q p is the flow rate per unit area of ​​the flow field.

[0068] On the one hand, the constant pressure specific heat capacity C p The influence of temperature is small; on the other hand, the convective heat transfer coefficient is mainly affected by the Reynolds number and is not greatly affected by small temperature changes. The Reynolds number is a function of the flow rate, so it can be written as:

[0069]

[0070] In the formula, f(q p ) is the convective heat transfer coefficient h with the flow rate per unit area q p Function of change.

[0071] The model can be further simplified as follows:

[0072]

[0073] Taking the scaling problem into consideration, the heat transfer problem will change: when scaling occurs in the heat exchanger, the scale adheres to the heat exchange wall surface. At this time, the heat transfer process changes from heat transfer from the fluid at the wall surface to convection heat transfer between the fluid and the scale at the scale surface, and then heat conduction from the scale surface to the heat exchange wall surface, then the above formula becomes:

[0074]

[0075] Where, T f is the surface temperature of the dirt. Since the thermal conductivity of dirt is much lower than that of the metal wall, when a scaling failure occurs, even a small thickness of scaling will cause a large change in temperature. Therefore, it can be considered that the surface temperature of the dirt is less affected by a small temperature change, but is greatly affected by scaling.

[0076] 2.3) Construction of following characteristic analysis indicators. The EG two-step method is introduced to analyze the following characteristics. Since the premise of cointegration analysis is that the time series used must be first-order single integration, only temperature and flow are selected as cointegration analysis indicators. Ordinary least squares regression calculation:

[0077] Y t =α+βX t +ε t

[0078] In the formula, variable X t is the hot side / cold measurement inlet temperature, Y tis the hot side / cold outlet temperature, which are all first-order single integral time series; α is the intercept of the equation; β is the slope of the regression equation, ε t is a random error term.

[0079] The regression equation is obtained:

[0080]

[0081] Subtracting the two equations, we get:

[0082]

[0083] in, is the estimated random error term; For Y t Estimated value of is the estimated value of α; is the estimated value of β.

[0084] ADF stationarity test is used to verify If it is stable, then X t and Y t Cointegration. The following characteristic expression of temperature can be obtained:

[0085]

[0086] In summary, the following characteristic analysis indicators are constructed as follows:

[0087]

[0088] in, is the estimated random error term in the heat flow temperature equation at time t; T ho,t is the heat flow outlet temperature at time t; is the estimated intercept in the heat flow temperature equation; is the estimated slope in the heat flow temperature equation; T hi,t is the heat flow inlet temperature at time t; is the estimated random error term in the cold flow temperature equation at time t; T co,t is the cold flow outlet temperature at time t; is the estimated intercept in the cold flow temperature equation; is the estimated slope in the cold flow temperature equation; T ci,t is the cold stream inlet temperature at time t.

[0089] In this embodiment, an abnormality detection model under non-steady-state multi-operating conditions is constructed, including: constructing a healthy memory matrix of multiple operating conditions, and simplifying the healthy memory matrix using the spatial correlation characteristics between matrices, so as to realize multi-operating condition, non-steady-state online abnormality detection of heat exchangers.

[0090] Specifically, the construction method of the anomaly detection model under non-steady-state multi-operating conditions is as follows:

[0091] Assume that there is a set of time series samples, which contains a total of m samples from time i to i+m, and there are e observations in each sample, which is defined as a healthy memory matrix D(t) with a length of m and a dimension of e:

[0092]

[0093] In the formula, a e (t) is the e-th observation at time t.

[0094] At the same time, there is an observation vector at time j, and each sample has an observation time series of e observations. Similarly, it can be recorded as X obs :

[0095]

[0096] In the formula, a e (j) is the e-th observation vector at time j.

[0097] Since the length of the observation vector is different from that of the time series sample matrix, in order to evaluate the relationship between them, the original time series matrix can be reduced to 1 through some linear transformation, so a set of weight coefficients W is introduced:

[0098] W=[w1,w2,w3…w e ] T

[0099] So that:

[0100]

[0101] Where, X est is the estimated vector. The essence of the estimated vector is the linear combination of the m normal states in the health memory matrix D. The weight vector W is the weight coefficient of the corresponding combination, which can be obtained by X obs With X est is determined by the residual between .

[0102] According to the estimated value X in the same period under healthy conditions est With real-time test value X obs Based on this principle, the healthy memory matrix can be used as the observation matrix, and the estimated distance is calculated for itself to obtain the benchmark estimated distance sequence:

[0103] dist base =[dist1,dist2...dist t ]

[0104]

[0105] And further calculate the mean and variance characteristics of the benchmark, denoted as μ base and σ base , the 3σ principle is introduced to identify abnormalities of observation vectors.

[0106] dist t >μ base +3σ base

[0107] When the above formula is satisfied, the observation vector is considered to be an abnormal observation vector. Since there are multiple operating conditions for the cooling water heat exchanger of the equipment, it is necessary to construct multiple sets of memory matrices to form a z-layer healthy memory matrix with e dimensions and m length. The distance between the observation vector and the entire memory matrix is ​​estimated. If there is no memory matrix that makes the above formula invalid, the vector is considered to deviate from any known healthy condition and is considered abnormal.

[0108] 3) Quantitative diagnosis of leakage faults and scaling faults is performed based on the abnormal detection model under non-steady-state multiple working conditions.

[0109] 3.1) Quantitative diagnosis of leakage faults, including: from the perspective of heat transfer and energy conservation, constructing a quantitative diagnosis model for internal leakage of the medium and flange sealing surface leakage based on heat balance to perform quantitative diagnosis of medium leakage.

[0110] In this embodiment, the construction of the quantitative diagnosis model for medium internal leakage and flange sealing surface leakage includes the following steps:

[0111] (1) Construction of quantitative diagnostic model for internal leakage of media:

[0112] The heat release Q1 of the leaked part of the fluid when it is not leaking is:

[0113] Q1=xcρ(T hi -T ho )

[0114] Where x is the leakage; c is the constant pressure specific heat of the cold fluid; ρ is the fluid density; T hi is the tube inlet temperature; T ho is the tube outlet temperature.

[0115] Assume that the temperature of the leaking hot fluid in the tube side and the cold fluid in the shell side is (T ci +T co ) / 2, the actual leakage amount Q2 of the leaking fluid is:

[0116]

[0117] Where, T ciis the shell inlet temperature, T co is the shell outlet temperature.

[0118] External heat Q caused by external factors (such as operating conditions fluctuations) causing the shell side cooling fluid temperature rise 外 for:

[0119] Q 外 =V c cρΔT ci

[0120] Where V c is the cold fluid flow rate.

[0121] The actual energy change Q3 of the cold fluid is:

[0122] Q3=V c cρΔT co

[0123] The final energy change of the cold fluid is caused by the change in the cold side inlet temperature and the transition of the leaked fluid from normal heat exchange to abnormal heat exchange. According to the law of heat conservation, we can get:

[0124] Q 外 +Q2=Q3+Q1

[0125] In the formula, the left side is the heat provided by the heat source, and the right side is the heat absorbed from the heat source. Simplified:

[0126]

[0127] The quantitative diagnosis model of internal leakage can be obtained. co is the cold flow outlet temperature difference; ΔT ci is the cold stream inlet temperature difference.

[0128] (2) Construction of quantitative diagnosis model for flange sealing surface leakage:

[0129] Flange leakage is a medium leakage, and the hot fluid in the tube side leaks from the flange sealing surface to the outside of the heat exchanger at the tube side inlet. At this time, the remaining hot fluid in the tube side of the heat exchanger exchanges heat with the cold fluid in the shell side, which can be expressed as:

[0130] (V h -x)(T hi -T ho )=V c (T co -T ci )

[0131] Where x is the leakage, V h is the thermal fluid flow rate.

[0132] The left side of the equal sign is the heat released by the tube side hot fluid after the leakage to the outside is eliminated, and the right side of the equal sign is the heat absorbed by the shell side cold fluid from the remaining hot fluid. Simplified:

[0133]

[0134] The quantitative diagnosis model of flange sealing surface leakage can be obtained.

[0135] 3.2) Quantitative diagnosis of scaling faults, including: constructing a scaling estimation method based on a benchmark heat transfer model, introducing a nonlinear regression model to construct a quantitative method for scaling thickness, and constructing a quantitative diagnostic model for scaling based on the response data of the actual system.

[0136] In this embodiment, the construction of the quantitative diagnostic model of scaling amount includes the following steps:

[0137] (1) Benchmark heat transfer model

[0138] The benchmark heat transfer model is to evaluate the heat transfer capacity of the heat exchanger under the current working conditions without scaling. Without considering other additional thermal resistances, the heat transfer coefficient K value for the single-layer enclosure structure can be calculated as follows:

[0139]

[0140] Among them, R h , R c , R w The thermal resistances represent the three processes of cold side convection heat transfer, hot side convection heat transfer and wall heat transfer, which are the reciprocals of the convection heat transfer coefficient or thermal conductivity of the three processes:

[0141]

[0142] In the formula, δ w and λ w They are the thickness and thermal conductivity of the heat transfer wall, which are mainly determined by the structure and material of the heat exchanger; α h and α c are the convective heat transfer coefficients on the hot side and the cold side, respectively. According to the Nusselt number and the forced convection heat transfer formula, the convective heat transfer coefficient can be expressed as:

[0143]

[0144] Where Re is the Reynolds number; Pr is the Prandtl number; L is the characteristic length; λ is the thermal conductivity of the fluid; α is the convective heat transfer coefficient.

[0145] The calculation formulas for Re and Pr are as follows:

[0146]

[0147] Where μ is the dynamic viscosity coefficient; C p is the specific heat at constant pressure; v and L are the characteristic velocity and equivalent length of the flow field, respectively; ρ is the fluid density. Substituting the formulas for Reynolds number and Prandtl number into the formulas, we get:

[0148]

[0149] For a specific heat exchanger, the standard expression of Reynolds number can be replaced by mass flow rate q and medium flow area s to obtain:

[0150] α=(CL m-1 c p n s -m )·λ 1-n ·μ n-m ·q m

[0151] Assuming that the structures on both sides of the heat exchanger are consistent and the physical and chemical properties of the fluids are similar, the formulas for the convective heat transfer coefficients on the cold and hot sides are as follows:

[0152]

[0153] The heat transfer coefficient K value can be expressed as:

[0154]

[0155] Since the thermal resistance R is the reciprocal of the heat transfer coefficient K, it can be written in the form of thermal resistance, that is:

[0156]

[0157] Since the calculated heat transfer coefficient does not take into account other additional thermal resistance, when fouling occurs in the heat exchanger, the heat transfer coefficient is changed to:

[0158]

[0159] In the formula, R f is the thermal resistance of scaling.

[0160] Substituting thermal resistance for heat transfer coefficient yields:

[0161]

[0162] In summary, in order to obtain the current fouling thermal resistance of the heat exchanger, the values ​​of the unknown parameters in the formula need to be obtained. According to the description of the modified Wilson method, when the flow rate on one side of the heat exchanger changes and the flow rate on the other side is fixed, the heat transfer criterion equation on that side can be determined, and then its unknown parameters such as C, m, and n can be determined. Therefore, for a heat exchanger with forced convection heat transfer, when its flow rate variation range tends to a limited range, its C, m, n and other parameters are objectively fixed values. Based on this assumption, since the heat exchanger characteristic length, constant pressure specific heat, flow area, wall thermal conductivity and wall thickness and other parameters are known constants, the fouling thermal resistance value can be further changed to:

[0163] R C =C1·λ h n-1 ·μ h m-n ·q h -m +C2·λ c n-1 ·μ c m-n ·q c -m +C3

[0164] In the formula, C1 and C2 are unknown constants. Move the thermal resistance value in the above formula to the right and we get:

[0165] d=C1·λ h n-1 ·μ h m-n ·q h -m +C2·λ c n-1 ·μ c m-n ·q c -m +C3-R C

[0166] In the formula, d is the error, that is, the difference between the fitting result and the actual value. According to the idea of ​​least squares method, the calculation of the trained model on the training set requires the error d to be minimum, that is:

[0167]

[0168] d i =y i -f(x)

[0169] Based on historical data such as the average flow rate, average viscosity, and average thermal conductivity of the cold and hot sides, the least squares method is introduced and the matrix solution is used to calculate the hyperparameter group with the minimum L. At this time, by inputting the monitoring parameters after the fault and combining the hyperparameter group of the benchmark heat transfer model, the healthy heat transfer coefficient under the fault condition can be directly calculated.

[0170] (2) Quantitative diagnosis model for scaling failure

[0171] However, the fouling thermal resistance cannot directly express the amount of fouling. The thermal conductivity of the fouling is estimated by using multiple sets of fouling tests, and then the fouling thickness is calculated. The expression of fouling thickness is as follows:

[0172] δ f =R f ·λ f

[0173] In the formula, δ f is the thickness of the dirt layer; R f is the thermal resistance of the dirt layer; w is the thermal conductivity of the dirt layer.

[0174] In one embodiment of the present invention, a heat exchanger abnormality detection and quantitative diagnosis system under non-steady-state multi-operating conditions is provided, which comprises:

[0175] The data processing module collects data of key components of the heat exchanger and pre-processes the data, where the data includes online monitoring data and offline data;

[0176] The following characteristic module constructs a fluctuation condition sensitive index based on the preprocessed data, and obtains the following characteristics of the online monitoring data from the fluctuation condition sensitive index to perform anomaly detection under non-stationary multi-conditions;

[0177] The diagnostic module performs quantitative diagnosis of leakage faults and scaling faults based on the test results.

[0178] In the above embodiment, the online monitoring parameters include temperature, pressure and volume flow rate; the offline data include structural parameters and physical parameters, wherein the structural parameters include the equivalent diameter of the cold / hot side flow channel, the flow channel cross-sectional area, the flow channel length, the number of processes, the total heat transfer area, and the pipe diameter ratio; the physical parameters include the absolute viscosity, density and constant pressure specific heat of the cold / hot side medium.

[0179] In the above embodiment, a fluctuation condition sensitive index is constructed based on the preprocessed data, and a follow-up characteristic analysis index of the online monitoring data is obtained from the fluctuation condition sensitive index, including:

[0180] Assume that the heat exchanger inlet parameter is C2 and the outlet parameter is C1; the heat exchanger outlet parameter C2 changes with the change of the inlet parameter C1, and the specific change is determined by the hidden state of the heat exchanger;

[0181] When the hidden state changes, the heat exchanger deviates from the original working state and an abnormality occurs; the following characteristics of the parameters reflect the hidden state of the heat exchanger. By analyzing the following characteristics of the monitoring parameters, the hidden state of the heat exchanger can be inferred, thereby achieving abnormality detection.

[0182] In the above embodiment, constructing an abnormality detection model under non-steady-state multiple working conditions includes:

[0183] A health memory matrix of multiple operating conditions is constructed, and the spatial correlation characteristics between matrices are used to simplify the health memory matrix to achieve multi-operating condition and non-stationary online anomaly detection of heat exchangers.

[0184] In the above embodiment, the quantitative diagnosis of leakage faults includes: starting from the perspective of heat transfer and energy conservation, constructing a quantitative diagnosis model of medium internal leakage and flange sealing surface leakage based on heat balance to perform quantitative diagnosis of medium leakage.

[0185] In the above embodiment, the quantitative diagnosis model for medium internal leakage and flange sealing surface leakage is:

[0186]

[0187] Where x is the leakage; V h is the thermal fluid flow rate; V c is the cold fluid flow rate; T hi is the tube inlet temperature; T ho is the tube outlet temperature, T ci is the shell inlet temperature, T co is the shell outlet temperature.

[0188] In the above embodiment, the quantitative diagnosis of scaling faults includes: constructing a scaling amount estimation method based on a reference heat transfer model, introducing a nonlinear regression model to construct a quantitative method for scaling thickness, and constructing a quantitative diagnostic model for scaling amount in combination with the response data of the actual system, which is:

[0189] δ f =R f ·λ f

[0190] In the formula, δ f is the thickness of the dirt layer; R f is the thermal resistance of the dirt layer; w is the thermal conductivity of the dirt layer.

[0191] The system provided in this embodiment is used to execute the above-mentioned method embodiments. Please refer to the above-mentioned embodiments for specific processes and detailed contents, which will not be repeated here.

[0192] In summary, the present invention introduces the idea of ​​cointegration analysis, proposes a multivariate state estimation anomaly detection method based on following characteristics, and constructs a health memory matrix of multiple working conditions by studying the correlation between operating condition fluctuations and state changes during the normal operation of the heat exchanger. The health memory matrix is ​​simplified by using the spatial correlation characteristics between matrices to achieve multi-condition, non-stationary online anomaly detection of heat exchangers. Finally, the qualitative fault identification and location technology and the quantitative fault degree analysis technology are combined to establish a heat exchanger fault diagnosis model based on the combination of qualitative and quantitative methods. The classification and location of faults are achieved by establishing a qualitative fault diagnosis expert system, and then the failure degree quantification method is constructed by deducing the mechanism of the failure mode to achieve accurate diagnosis of complex failures.

[0193] In one embodiment of the present invention, a computing device is provided, which may be a terminal, and may include: a processor, a communication interface, a memory, a display screen, and an input device. Among them, the processor, the communication interface, and the memory communicate with each other through a communication bus. The processor is used to provide computing and control capabilities. The memory includes a non-volatile storage medium and an internal memory, and the non-volatile storage medium stores an operating system and a computer program. When the computer program is executed by the processor, the method in each of the above embodiments is implemented; the internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a management network, NFC (near field communication) or other technologies. The display screen may be a liquid crystal display screen or an electronic ink display screen, and the input device may be a touch layer covered on the display screen, or a key, trackball or touchpad provided on the housing of the computing device, or an external keyboard, touchpad or mouse, etc. The processor may call the logic instructions in the memory.

[0194] In addition, the logic instructions in the above-mentioned memory can be implemented in the form of software functional units and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0195] In one embodiment of the present invention, a computer program product is provided, wherein the computer program product includes a computer program stored on a non-transitory computer-readable storage medium, wherein the computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the methods provided by the above-mentioned method embodiments.

[0196] In one embodiment of the present invention, a non-transitory computer-readable storage medium is provided, wherein the non-transitory computer-readable storage medium stores server instructions, and the computer instructions enable a computer to execute the methods provided in the above embodiments.

[0197] The above embodiment provides a computer-readable storage medium, whose implementation principle and technical effect are similar to those of the above method embodiment, and will not be repeated here.

[0198] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0199] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0200] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0201] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting and quantifying abnormalities in a heat exchanger under non-steady-state conditions, characterized in that: include: Collect data of key components of heat exchangers and pre-process the data, where the data of key components includes online monitoring data and offline data; Based on the preprocessed data, the fluctuating condition sensitive indicators are constructed. The fluctuating condition sensitive indicators include leakage fault indicators and scaling fault indicators. The following characteristic analysis indicators of the online monitoring data are obtained from the fluctuating condition sensitive indicators to construct an abnormality detection model under non-steady-state multi-conditions. Quantitative diagnosis of leakage faults and scaling faults is performed based on the abnormal detection model under non-steady-state multiple working conditions.

2. The method for detecting and quantifying abnormalities in a heat exchanger under non-steady-state conditions as claimed in claim 1, characterized in that: Online monitoring parameters include temperature, pressure and volume flow; Offline data includes structural parameters and physical parameters. The structural parameters include the equivalent diameter of the cold / hot side flow channel, the flow channel cross-sectional area, the flow channel length, the number of processes, the total heat transfer area, and the pipe diameter ratio; the physical parameters include the absolute viscosity, density and constant pressure specific heat of the cold / hot side medium.

3. The method for detecting and quantifying abnormalities in a heat exchanger under non-steady-state conditions as claimed in claim 1, characterized in that: Based on the preprocessed data, a fluctuation condition sensitive index is constructed, and the following characteristic analysis index of the online monitoring data is obtained from the fluctuation condition sensitive index, including: Assume that the heat exchanger inlet parameter is C2 and the outlet parameter is C1; the heat exchanger outlet parameter C2 changes with the change of the inlet parameter C1, and the specific change is determined by the hidden state of the heat exchanger; When the hidden state changes, the heat exchanger deviates from the original working state and an abnormality occurs; the following characteristics of the parameters reflect the hidden state of the heat exchanger. By analyzing the following characteristics of the monitoring parameters, the hidden state of the heat exchanger can be inferred, thereby achieving abnormality detection.

4. The method for detecting and quantifying abnormalities in a heat exchanger under non-steady-state conditions as claimed in claim 3, characterized in that: Construct anomaly detection models under non-steady-state multi-operating conditions, including: A health memory matrix of multiple operating conditions is constructed, and the spatial correlation characteristics between matrices are used to simplify the health memory matrix to achieve multi-operating condition and non-stationary online anomaly detection of heat exchangers.

5. The method for detecting and quantifying abnormalities in a heat exchanger under non-steady-state conditions as claimed in claim 1, characterized in that: Quantitative diagnosis of leakage faults includes: from the perspective of heat transfer and energy conservation, building a quantitative diagnosis model for internal leakage of the medium and flange sealing surface leakage based on heat balance to perform quantitative diagnosis of medium leakage.

6. The method for detecting and quantifying abnormalities in a heat exchanger under non-steady-state multi-operating conditions as claimed in claim 1, characterized in that: The quantitative diagnosis model for internal leakage of media and leakage of flange sealing surface is: Where x is the leakage; V h is the thermal fluid flow rate; V c is the cold fluid flow rate; T hi is the tube inlet temperature; T ho is the tube outlet temperature, T ci is the shell inlet temperature, T co is the shell outlet temperature.

7. The method for detecting and quantifying abnormalities in a heat exchanger under non-steady-state conditions as claimed in claim 1, characterized in that: Quantitative diagnosis of scaling faults includes: constructing a scaling estimation method based on a benchmark heat transfer model, introducing a nonlinear regression model to construct a quantitative method for scaling thickness, and constructing a quantitative diagnostic model for scaling in combination with the response data of the actual system, which is: d f =R f ·l f In the formula, δ f is the thickness of the dirt layer; R f is the thermal resistance of the dirt layer; w is the thermal conductivity of the dirt layer.

8. A heat exchanger abnormality detection and quantitative diagnosis system under non-steady-state multi-operating conditions, characterized in that: include: The data processing module collects data of key components of the heat exchanger and pre-processes the data, where the data includes online monitoring data and offline data; The following characteristic module constructs the fluctuation condition sensitive index based on the preprocessed data. The fluctuation condition sensitive index includes leakage fault index and scaling fault index. The following characteristic analysis index of the online monitoring data is obtained from the fluctuation condition sensitive index to construct the abnormality detection model under non-steady-state multi-conditions. The diagnostic module performs quantitative diagnosis of leakage faults and scaling faults based on the abnormal detection model under non-steady-state multiple working conditions.

9. A computer-readable storage medium storing one or more programs, characterized in that: The one or more programs include instructions, which, when executed by a computing device, cause the computing device to perform any one of the methods of claims 1 to 7.

10. A computing device, characterized in that include: One or more processors, a memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, and the one or more programs include instructions for executing any one of the methods described in claims 1 to 7.

Images