Remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion

Abstract

The present invention discloses a remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion, which relates to the technical field of environmental protection equipment fault diagnosis. The system comprises the following steps: installing a sensor network at key locations in a plastic pyrolysis reactor and storing the collected signals in a historical reaction database; performing signal decomposition on all data in the historical reaction database to obtain an environmental disturbance component and a catalyst memory component, inputting these components into a constructed component prediction model to obtain a catalyst memory influence coefficient and an environmental disturbance influence coefficient; obtaining a reactor risk index based on the catalyst memory residual and the environmental disturbance residual in combination with temperature deviation; determining whether to operate a catalyst difference data module through a cleaning judgment strategy; and calculating the catalyst-temperature influence coefficient after cleaning and transmitting it to the component prediction model. The present invention solves the problem of misjudgment caused by process fluctuations and the catalyst "memory" effect, thereby improving the operational safety, monitoring accuracy, and operation and maintenance efficiency of the plastic pyrolysis reactor.

Description

Technical Field

[0001] The present invention relates to the technical field of environmental protection equipment fault diagnosis, and in particular to a remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion. Background Art

[0002] With the rapid development of the new energy industry, waste battery recycling has become a key link in the circular economy and environmental protection. The plastic separators and packaging materials of waste lithium-ion batteries are recycled through pyrolysis technology in plastic pyrolysis reactors. This not only removes harmful binders and electrolyte residues, but also converts the plastic into fuel oil and combustible gas, realizing resource utilization. Therefore, high-precision fault monitoring of plastic pyrolysis reactors is of great significance for improving the operational reliability and intelligence level of pyrolysis equipment and promoting the sustainable development of the waste plastic resource industry.

[0003] For example, the existing Chinese patent publication number CN110208451B discloses a bipolar micro-fixed bed reactor combined with a photoionization mass spectrometry online detection system and method, including a bipolar micro-fixed bed reactor, an offline cooling device, and an online detection device, wherein the bipolar micro-fixed bed reactor is respectively connected to the offline cooling device and the online detection device, and the bipolar micro-fixed bed reactor is connected to the offline cooling device via a transmission line; the bipolar micro-fixed bed reactor catalytically pyrolyzes a solid fuel sample; the pyrolysis reaction and the catalytic reaction of the present invention do not interfere with each other, the reactor and the heating unit can be quickly separated and cooled, and two different types of coke in a primary state can be accurately analyzed; the two reactors are separated by a high-temperature valve, which can realize intermittent continuous sampling of solid fuel and real-time monitoring of catalyst deactivation, and the device and reactor are small in size and the residence time of the reactants is short;

[0004] However, for traditional plastic pyrolysis reactors, the high temperature, oxygen-free environment and the adsorption-desorption behavior of cracking products on the catalyst surface cause the acidic products adsorbed in the equipment to shift the acid-base balance on the catalyst surface toward the acid side. The activation energy required for cracking changes accordingly, causing the rate constant to drift at the same temperature, resulting in a "history dependence" effect and transient disturbances inside the reactor, which can cause false alarms and bring major challenges to equipment operation monitoring and fault diagnosis.

[0005] To this end, the present invention provides an environmental protection equipment remote intelligent fault diagnosis system based on multi-source data fusion. Summary of the Invention

[0006] The purpose of the present invention is to provide a remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion to solve the existing problems raised in the above background technology.

[0007] To achieve the above objectives, the present invention provides the following technical solutions: a remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion, comprising:

[0008] Signal acquisition module, used to install sensor networks at key locations in the plastic pyrolysis reactor and store the collected signals in a historical reaction database;

[0009] The signal decomposition module is used to decompose all the data in the historical reaction database to obtain the environmental disturbance component and the catalyst memory component, and input them into the constructed component prediction model to obtain the catalyst memory influence coefficient and the environmental disturbance influence coefficient;

[0010] The safety threat prediction module obtains the reactor risk index based on the catalyst memory residual and the environmental disturbance residual combined with the temperature deviation; and determines whether to run the catalyst difference data module through the cleaning judgment strategy;

[0011] The catalyst difference data module is used to calculate the catalyst-temperature influence coefficient after cleaning and transmit it to the component prediction model.

[0012] A further improvement of the present invention is that the signal decomposition module includes a signal separation unit, a component prediction model construction unit and an adaptive threshold generation unit;

[0013] The signal separation unit is equipped with an adsorption-desorption critical frequency calculation strategy. First, the adsorption-desorption time constant is determined. In the catalytic cracking reaction, the adsorption-desorption of the product molecule A on the catalyst surface S is expressed as ,in, is the adsorption rate constant, represents the desorption rate constant; then its characteristic time constant is expressed as ,in, represents the instantaneous concentration of the adsorbate in the gas phase of the reactor, and the corresponding adsorption-desorption critical frequency is expressed as .

[0014] The present invention is further improved in that the signal separation unit further includes the original signal Apply complete set empirical mode decomposition to obtain N intrinsic mode function components , for each Perform Hilbert transform to get the instantaneous frequency , get the average frequency of each mode ,like If it is greater than k, the corresponding component is extracted and classified as the environmental disturbance component. ,like If it is less than or equal to k, the corresponding component is extracted and classified as the catalytic memory component .

[0015] A further improvement of the present invention is that the component prediction model construction unit includes aligning the environmental disturbance component and the catalyst memory component output by the signal decomposition unit according to timestamps to form a parallel time series sequence; for each time t, intercepting the forward input with a window length w, and normalizing the data within the window;

[0016] Then, an environmental disturbance channel and a catalytic memory channel are constructed. The environmental disturbance channel performs causal and dilated convolution on the environmental disturbance items of the parallel time series through the first layer of convolution kernel of the temporal convolution network, outputs the environmental disturbance feature map, and the second layer of convolution kernel convolves the output of the first layer. This is repeated until the Mth layer of convolution kernel, and the M feature maps are compressed into a vector of fixed dimension. , each layer of the decoder performs a deconvolution operation, recovering the sequence length and channel layer by layer, and generating the same as the input data at the last layer Sequences of the same dimension , to achieve the reconstruction of the environmental disturbance component, and The root mean square error is used as the reconstruction loss and calculated and The difference output of the environmental disturbance residual is used as the environmental disturbance influence coefficient.

[0017] A further improvement of the present invention is that the catalytic memory channel is implemented based on the LSTM model, specifically including:

[0018] Computes the sliding mean of the catalytic memory term of a parallel time series and autocorrelation coefficient ,Will 、 The feature dimension is spliced with the original catalytic memory component to form a three-channel catalytic memory input tensor Cmi; the LSTM model accepts these three sequences at each step and performs normalization, mapping the last layer of full connection to the scalar prediction to obtain the predicted value of the catalytic memory component at time t+1. , and then get ,calculate and The difference between the two is used to output the catalytic memory residual as the catalytic memory influence coefficient.

[0019] The present invention is further improved in that the adaptive threshold generation unit includes extracting the adsorption-desorption critical frequency k and the environmental disturbance influence coefficient and catalytic memory influence coefficient Perform wavelet packet transform to obtain the energy of each subband, and calculate the ratio of the sum of the energy corresponding to the environmental disturbance term and the catalytic memory term to obtain the catalytic memory energy ratio ; The adsorption-desorption critical frequency k and catalytic memory energy ratio The fuzzy control input is obtained by the tanh activation function and , get the environmental interference threshold and catalytic memory threshold ,in is the comprehensive scaling coefficient output by fuzzy reasoning, represents the environmental disturbance influence coefficient within the sliding window [t−w,t], represents the catalytic memory influence coefficient within the sliding window [t−w,t].

[0020] The present invention is further improved in that the security threat prediction module includes a temperature deviation calculation unit and an equipment risk acquisition unit; the temperature deviation calculation unit calculates the device risk by measuring the temperature. Predicting Temperature with Digital Twins The temperature deviation is obtained by the difference The equipment risk acquisition unit obtains the reactor threat fusion value by standardizing the catalyst memory influence coefficient, the environmental disturbance influence coefficient and the temperature deviation, and then uses the Euclidean norm aggregation to obtain the reactor threat fusion value, and maps the reactor threat fusion value to [0,1] through the Logistic function to obtain the reactor risk index. .

[0021] A further improvement of the present invention is that the equipment risk acquisition unit is further equipped with a cleaning judgment strategy, which is implemented by setting a cleaning risk threshold Tcl and an alarm risk threshold Tal;

[0022] When the reactor risk index When the value is less than the cleaning risk threshold Tcl, the alarm level is normal and monitoring continues without any operation.

[0023] When the reactor risk index When it is greater than or equal to the cleaning risk threshold Tcl and less than the alarm risk threshold Tal, the alarm level is warning. Automatically adjust the flow rate and arrange online ultrasonic or air blowing cleaning, Indicates the initial flow rate; simultaneously runs the catalyst difference data module;

[0024] When the reactor risk index When the value is greater than the alarm risk threshold Tal, the alarm level is critical, catalyst oxidation regeneration or safe shutdown is started, a maintenance work order is triggered, and an alarm notification is sent.

[0025] A further improvement of the present invention is that the catalyst difference data module includes recording the temperature deviation of the digital twin before cleaning at the cleaning trigger moment. , after cleaning is completed and operation is resumed, record the new deviation , and obtain the deviation recovery amount ,Will and the catalytic memory component in the corresponding time window Pairing to form an incremental learning sample set, adding new samples to the LSTM training queue, and using a small batch model for training and the catalytic memory component within the sliding window [t−w,t] , and obtain the catalyst-temperature influence coefficient , after splicing the feature dimensions, update the model weights.

[0026] A further improvement of the present invention is that the key positions of the plastic pyrolysis reactor include thermocouple arrays equidistantly arranged along the height direction inside the reactor cylinder and on the outer wall to measure the local temperature; and in-situ spectral probes are installed above the catalyst bed or at the sampling hole to detect the catalyst surface coverage.

[0027] Compared with the prior art, the present invention has the following beneficial effects:

[0028] The present invention firstly compares the adsorption-desorption critical frequency with the average instantaneous frequency of each signal component to accurately distinguish between environmental disturbances and catalytic memory effects, thus reducing the "slow drift" misjudgment caused by memory accumulation.

[0029] The signal decomposition module is trained on “normal” environmental disturbances to accurately reconstruct common patterns, producing significant residuals only for true anomalies, including sudden disturbances or equipment failures, thereby reducing the reconstruction-based false alarm rate;

[0030] Introducing the high / low frequency energy ratio in adaptive threshold generation can dynamically adjust the sensitivity according to the energy distribution, making the system robust to slow-changing memory noise and highly alert to transient disturbances;

[0031] Thresholds are dynamically generated based on the statistical characteristics of the residual window, which can track changes in operating conditions in real time and automatically adjust the threshold bandwidth, significantly reducing false alarms caused by process drift; multi-source information such as adsorption driving factors, energy ratios and temperature deviations are integrated through Mamdani reasoning to effectively balance the adjustment of thresholds in the "memory accumulation" and "sudden disturbance" scenarios, ensuring the optimal balance between sensitivity and stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a framework diagram of the environmental protection equipment remote intelligent fault diagnosis system based on multi-source data fusion of the present invention;

[0033] Figure 2 This is a flow chart of the security threat prediction module of the present invention. DETAILED DESCRIPTION

[0034] The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present invention and the specific features in the embodiments are detailed descriptions of the technical solution of the present invention, rather than limitations on the technical solution of the present invention. In the absence of conflict, the embodiments of the present invention and the technical features in the embodiments can be combined with each other.

[0035] The term "and / or" simply describes a relationship between related objects, indicating that three possible relationships exist. For example, "A and / or B" can mean: A exists alone, A and B exist simultaneously, or B exists alone. Additionally, the character " / " generally indicates an "or" relationship between the related objects.

[0036] Example 1

[0037] Figure 1 The framework diagram of the remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion disclosed in this embodiment is shown, including:

[0038] Signal acquisition module, used to install sensor networks at key locations in the plastic pyrolysis reactor and store the collected signals in a historical reaction database;

[0039] The signal decomposition module is used to decompose all the data in the historical reaction database to obtain the environmental disturbance component and the catalyst memory component, and input them into the constructed component prediction model to obtain the catalyst memory influence coefficient and the environmental disturbance influence coefficient;

[0040] The safety threat prediction module obtains the reactor risk index based on the catalyst memory residual and the environmental disturbance residual combined with the temperature deviation; and determines whether to run the catalyst difference data module through the cleaning judgment strategy;

[0041] The catalyst difference data module is used to calculate the catalyst-temperature influence coefficient after cleaning and output it to the component prediction model.

[0042] The key positions of the plastic pyrolysis reactor include a mass flowmeter at the feed port to measure the feed rate; upstream of the plastic particle / powder feed pipeline, between the inlet and the storage bin, thermocouples and mass flowmeters at key positions of the reactor collect data at a frequency of 1 Hz through the SCADA system and store them in a historical database; and thermocouple arrays (inlet, middle, outlet) are arranged equidistantly along the height direction of the inside and outer wall of the reactor cylinder, and local temperatures are measured at typically 5-7 measuring points; the full-field temperature distribution is calculated once a minute through the CFD-DEM digital twin and is synchronously aligned with the measured temperature; and an in-situ spectral probe is installed above the catalyst bed or at the sampling hole to detect the surface coverage of the bed.

[0043] Subsequently, bandpass filtering is applied to the original signal to remove electromagnetic interference of 50 Hz and above, and then linear detrending is performed; the temperature and flow signals are linearly mapped to the [0,1] interval to provide a unified dimension for subsequent decomposition and model training.

[0044] The signal decomposition module includes a signal separation unit, a component prediction model construction unit and an adaptive threshold generation unit;

[0045] The signal separation unit is equipped with an adsorption-desorption critical frequency calculation strategy. First, the adsorption-desorption time constant is determined. In the catalytic cracking reaction, the adsorption-desorption of the product molecule A on the catalyst surface S is expressed as ,in, is the adsorption rate constant, represents the desorption rate constant; it is set by those skilled in the art, and the secondary adsorption-desorption process is regarded as a first-order system, and its characteristic time constant is expressed as ,in, It represents the instantaneous concentration of adsorbate (cracking product) in the gas phase of the reactor, in units of , which can be obtained in real time by online gas chromatography or mass spectrometry;

[0046] Adsorption rate Refers to the amount of gas phase molecules A adsorbed per unit surface free site per unit time ,in, The unit is , is the concentration of free active sites per unit area or volume, in units ; Desorption rate , The unit is , represents the concentration of sites occupied by adsorbed molecules, and its unit is the same as Same; surface coverage Perform mass balance, represents the total site concentration, which describes the transient behavior of the adsorption-desorption process. , so that the equilibrium state ,have ,get ,in , represents the adsorption equilibrium constant, reflecting the affinity of the molecule on the surface;

[0047] The adsorption / desorption rate constants satisfy the Arrhenius relationship, , ,in and represents the pre-exponential factor, and represents activation energy, RT represents gas constant, and the respective and A; then correct T online to get real-time and ;exist Smaller or When it is larger, desorption is slow, adsorption is fast, and surface coverage is high. The higher the accumulation, the more significant the impact on the subsequent reaction rate, resulting in enhanced "history dependence", which can increase the characteristic frequency of adsorption-desorption. As the critical value for signal separation, the component with a frequency lower than this value is classified as the catalytic memory component, and the component with a frequency higher than this value is classified as the environmental disturbance component. The corresponding adsorption-desorption critical frequency is expressed as , representing the main dynamic time scale of the “memory” effect on the catalyst surface.

[0048] The signal separation unit also includes the original signal Apply complete set empirical mode decomposition to obtain N intrinsic mode function components , for each Perform Hilbert transform to get the instantaneous frequency , get the average frequency of each mode , ,like If it is greater than k, the corresponding component is extracted and classified as the environmental disturbance component. ,like If it is less than or equal to k, the corresponding component is extracted and classified as the catalytic memory component ; The Hilbert–Huang transform ensures accurate extraction of the local frequency of non-stationary and nonlinear signals.

[0049] The component prediction model construction unit includes aligning the environmental disturbance component and the catalyst memory component output by the signal decomposition unit according to the timestamp to form a parallel time series sequence; for each time t, intercepting the forward input with a window length w (for example, 100 steps), , , and standardize the data within the window;

[0050] Then, an environmental disturbance channel and a catalytic memory channel are constructed. The environmental disturbance channel performs causal and dilated convolution on the environmental disturbance items of the parallel time series through the first layer of convolution kernel of the temporal convolution network, outputs the environmental disturbance feature map, and the second layer of convolution kernel convolves the output of the first layer. This is repeated until the Mth layer of convolution kernel, and the M feature maps are compressed into a vector of fixed dimension. Each layer of the decoder performs a deconvolution operation, recovering the sequence length and channel layer by layer, so that the time dimension and number of channels of the feature map are restored in sequence, and the last layer generates the same as the input data. Sequences of the same dimension , to achieve the reconstruction of the environmental disturbance component, this vector is the "potential representation", which condenses all the time series information in the window. and The root mean square error is used as the reconstruction loss, , which measures the quality of reconstruction and drives the model to learn the “normal dynamics” of the perturbation. and The difference output of the environmental disturbance residual is used as the environmental disturbance influence coefficient.

[0051] The catalytic memory channel is implemented based on the LSTM model, specifically including

[0052] Computes the sliding mean of the catalytic memory term of a parallel time series and autocorrelation coefficient ,Will 、 The feature dimension is spliced with the original catalytic memory component to form a three-channel catalytic memory input tensor Cmi; the LSTM model accepts these three sequences at the same time in each step and performs normalization processing, so that it can refer to the original signal and auxiliary features when updating the hidden state, enhance the expression ability of trends and dependencies, and map the last layer of full connection to the scalar prediction to obtain the predicted value of the catalytic memory component at time t+1. , and then get ,calculate and The difference between the output and the catalytic memory residual is used as the catalytic memory influence coefficient, reflecting the cumulative "memory" effect not captured by the model.

[0053] By continuous calculation , you can get a ratio A much smoother curve is used to characterize the slow-changing trend and steady-state drift of the system; As a model input, it can provide a "local baseline" for the network to help it distinguish between true structural changes (trend drift) and random jitter (residuals), thereby reducing excessive sensitivity to short-term fluctuations;

[0054] A high positive autocorrelation (close to 1) means It changes slowly in adjacent moments and has a strong lag memory; while close to 0 indicates more random dynamics; calculated within the sliding window (e.g. window=50), which can quantify the "stableness" and "trend strength" of the window and provide a serial dependency indicator for the model;

[0055] Auxiliary features participate in weight updates in both network feedforward and backpropagation, allowing the model to focus more on low-frequency information related to trends rather than being dominated by short-term jitters; the sliding mean filters high-frequency fluctuations, and the autocorrelation value reflects the persistence of the signal. The two can complement each other, helping the model distinguish between "short-term disturbances" and "long-term trends" and reduce the false alarm rate.

[0056] The adaptive threshold generation unit includes extracting the adsorption-desorption critical frequency k. When the adsorption accumulation is significant (k is large), the threshold is enlarged to tolerate the "memory" drift; when k is small, the threshold is contracted to sensitively capture sudden disturbances and the influence coefficient of environmental disturbance is reduced. and catalytic memory influence coefficient Perform wavelet packet transform to obtain the energy of each subband, and calculate the ratio of the sum of the energy corresponding to the environmental disturbance term and the catalytic memory term to obtain the catalytic memory energy ratio ; A large value indicates that transient disturbances are prominent and the threshold should be appropriately lowered to improve sensitivity; hours can improve the threshold to robustly resist slow-changing noise; the adsorption-desorption critical frequency k and catalytic memory energy ratio The fuzzy control input is obtained by the tanh activation function and , get the environmental interference threshold and catalytic memory threshold ,in is the comprehensive scaling coefficient output by fuzzy reasoning, represents the environmental disturbance influence coefficient within the sliding window [t−w,t], represents the catalytic memory influence coefficient in the sliding window [t−w,t], and the fuzzy rule base is expressed as High and If the threshold is low, the Low and If it is high, the threshold is reduced, and the reasoning and deblurring adopt Mamdani reasoning, and deblurring is done by the centroid method; represents the mean value within the window, represents the standard deviation within the window.

[0057] Example 2

[0058] Based on the same inventive concept of Example 1, this embodiment proposes a specific implementation of the security threat prediction module in Example 1. Figure 2 The flow chart of the security threat prediction module of this embodiment is presented, which specifically includes a temperature deviation calculation unit and a device risk acquisition unit;

[0059] The temperature deviation calculation unit is based on the measured temperature Predicting Temperature with Digital Twins The temperature deviation is obtained by the difference The equipment risk acquisition unit obtains the reactor threat fusion value by standardizing the catalyst memory influence coefficient, the environmental disturbance influence coefficient and the temperature deviation, and then uses the Euclidean norm aggregation to obtain the reactor threat fusion value, and maps the reactor threat fusion value to [0,1] through the Logistic function to obtain the reactor risk index. .

[0060] The equipment risk acquisition unit is also equipped with a cleaning judgment strategy, which is implemented by setting a cleaning risk threshold Tcl and an alarm risk threshold Tal;

[0061] When the reactor risk index When the value is less than the cleaning risk threshold Tcl, the alarm level is normal and monitoring continues without any operation.

[0062] When the reactor risk index When it is greater than or equal to the cleaning risk threshold Tcl and less than the alarm risk threshold Tal, the alarm level is warning. Automatically adjust the flow rate and arrange online ultrasonic or air blowing cleaning, Indicates the initial flow rate; simultaneously runs the catalyst difference data module; It means that as the risk index of the reactor increases, the flow rate increases. It also increases accordingly, and the flow rate can be adaptively adjusted according to the risk index of the reactor;

[0063] When the reactor risk index When the value is greater than the alarm risk threshold Tal, the alarm level is critical, catalyst oxidation regeneration or safe shutdown is started, a maintenance work order is triggered, and an alarm notification is sent.

[0064] This embodiment also proposes a specific implementation method of the catalyst difference data module in Example 1, including recording the temperature deviation of the digital twin before cleaning at the cleaning trigger moment. , after cleaning is completed and operation is resumed, record the new deviation , and obtain the deviation recovery amount ,Will and the catalytic memory component in the corresponding time window Pairing to form an incremental learning sample set, adding new samples to the LSTM training queue, and using a small batch model for training and the catalytic memory component within the sliding window [t−w,t] , get the function f, calculate the catalyst-temperature influence coefficient After the feature dimension is spliced, the model weight is updated. Because the disturbance residual discrimination is more reliable, the overall false alarm rate of the system is reduced; the present invention can be based on Distribution automatically optimizes the cleaning cycle, when When the pressure is relatively low, the cleaning interval is extended, saving maintenance costs; the online learning mechanism enables the model to adapt to different raw material batches and operating conditions, improving system stability.

[0065] The thresholds, weights and other setting values may be set by default according to the present invention, or may be set by those skilled in the art.

[0066] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0067] 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 flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, 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 flowcharts and / or block diagrams. 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.

[0068] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work 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 The function specified in one or more boxes.

[0069] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0070] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which are all protected by the present invention.

Claims

1. A remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion, characterized by: include: Signal acquisition module, used to install sensor networks at key locations in the plastic pyrolysis reactor and store the collected signals in a historical reaction database; The signal decomposition module is used to decompose all the data in the historical reaction database to obtain the environmental disturbance component and the catalyst memory component, and input them into the constructed component prediction model to obtain the catalyst memory influence coefficient and the environmental disturbance influence coefficient; The safety threat prediction module obtains the reactor risk index based on the catalyst memory residual and the environmental disturbance residual combined with the temperature deviation; and determines whether to run the catalyst difference data module through the cleaning judgment strategy; Catalyst difference data module, used to calculate the catalyst-temperature influence coefficient after cleaning and transmit it to the component prediction model; The signal decomposition module includes a signal separation unit, a component prediction model construction unit and an adaptive threshold generation unit; The signal separation unit is equipped with an adsorption-desorption critical frequency calculation strategy. First, the adsorption-desorption time constant is determined. In the catalytic cracking reaction, the adsorption-desorption of the product molecule A on the catalyst surface S is expressed as ,in, is the adsorption rate constant, represents the desorption rate constant; then its characteristic time constant is expressed as ,in, represents the instantaneous concentration of the adsorbate in the gas phase of the reactor, and the corresponding adsorption-desorption critical frequency is expressed as ; The signal separation unit also includes the original signal Apply complete set empirical mode decomposition to obtain N intrinsic mode function components , for each Perform Hilbert transform to get the instantaneous frequency , get the average frequency of each mode ,like If it is greater than k, the corresponding component is extracted and classified as the environmental disturbance component. ,like If it is less than or equal to k, the corresponding component is extracted and classified as the catalytic memory component .

2. The remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion according to claim 1 is characterized by: The component prediction model construction unit includes aligning the environmental disturbance component and the catalyst memory component output by the signal decomposition unit according to timestamps to form a parallel time series sequence; for each time t, intercepting the forward input with a window length w, and normalizing the data in the window; Then, an environmental disturbance channel and a catalytic memory channel are constructed. The environmental disturbance channel performs causal and dilated convolution on the environmental disturbance items of the parallel time series through the first layer of convolution kernel of the temporal convolution network, outputs the environmental disturbance feature map, and the second layer of convolution kernel convolves the output of the first layer. This is repeated until the Mth layer of convolution kernel, and the M feature maps are compressed into a vector of fixed dimension. , each layer of the decoder performs a deconvolution operation, recovering the sequence length and channel layer by layer, and generating the same as the input data at the last layer Sequences of the same dimension , to achieve the reconstruction of the environmental disturbance component, and The root mean square error is used as the reconstruction loss and calculated and The difference output of the environmental disturbance residual is used as the environmental disturbance influence coefficient.

3. The remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion according to claim 2 is characterized by: The catalytic memory channel is implemented based on the LSTM model, specifically including: Computes the sliding mean of the catalytic memory term of a parallel time series and autocorrelation coefficient ,Will 、 The feature dimension is spliced with the original catalytic memory component to form a three-channel catalytic memory input tensor Cmi; the LSTM model accepts these three sequences at each step and performs normalization, mapping the last layer of full connection to the scalar prediction to obtain the predicted value of the catalytic memory component at time t+1. , and then get ,calculate and The difference between the two is used to output the catalytic memory residual as the catalytic memory influence coefficient.

4. The remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion according to claim 3 is characterized by: The adaptive threshold generation unit includes extracting the adsorption-desorption critical frequency k, the environmental disturbance influence coefficient and catalytic memory influence coefficient Perform wavelet packet transform to obtain the energy of each subband, and calculate the ratio of the sum of the energy corresponding to the environmental disturbance term and the catalytic memory term to obtain the catalytic memory energy ratio ; The adsorption-desorption critical frequency k and catalytic memory energy ratio The fuzzy control input is obtained by the tanh activation function and , get the environmental interference threshold and catalytic memory threshold ,in is the comprehensive scaling coefficient output by fuzzy reasoning, represents the environmental disturbance influence coefficient within the sliding window [t−w,t], represents the catalytic memory influence coefficient within the sliding window [t−w,t].

5. The remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion according to claim 4 is characterized by: The security threat prediction module includes a temperature deviation calculation unit and an equipment risk acquisition unit; the temperature deviation calculation unit measures the temperature Predicting Temperature with Digital Twins The temperature deviation is obtained by the difference The equipment risk acquisition unit obtains the reactor threat fusion value by standardizing the catalyst memory influence coefficient, the environmental disturbance influence coefficient and the temperature deviation, and then uses the Euclidean norm aggregation to obtain the reactor threat fusion value, and maps the reactor threat fusion value to [0,1] through the Logistic function to obtain the reactor risk index. .

6. The remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion according to claim 5 is characterized by: The equipment risk acquisition unit is also equipped with a cleaning judgment strategy, which is implemented by setting a cleaning risk threshold Tcl and an alarm risk threshold Tal; When the reactor risk index When the value is less than the cleaning risk threshold Tcl, the alarm level is normal and monitoring continues without any operation. When the reactor risk index When it is greater than or equal to the cleaning risk threshold Tcl and less than the alarm risk threshold Tal, the alarm level is warning. Automatically adjust the flow rate and arrange online ultrasonic or air blowing cleaning, Indicates the initial flow rate; simultaneously runs the catalyst difference data module; When the reactor risk index When the value is greater than the alarm risk threshold Tal, the alarm level is critical, catalyst oxidation regeneration or safe shutdown is started, a maintenance work order is triggered, and an alarm notification is sent.

7. The remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion according to claim 6 is characterized by: The catalyst difference data module includes recording the temperature deviation of the digital twin before cleaning at the cleaning trigger moment. , after cleaning is completed and operation is resumed, record the new deviation , and obtain the deviation recovery amount ,Will and the catalytic memory component in the corresponding time window Pairing to form an incremental learning sample set, adding new samples to the LSTM training queue, and using a small batch model for training and the catalytic memory component within the sliding window [t−w,t] , and obtain the catalyst-temperature influence coefficient , after splicing the feature dimensions, update the model weights.

8. The remote intelligent fault diagnosis system for environmental protection equipment based on multi-source data fusion according to claim 7 is characterized by: The key positions of the plastic pyrolysis reactor include thermocouple arrays equidistantly arranged inside the reactor cylinder and on the outer wall along the height direction to measure local temperature; and in-situ spectral probes are installed above the catalyst bed or at the sampling hole to detect the catalyst surface coverage.

Images