A chemical plant chemical data analysis system and method based on industrial big data

Abstract

The invention belongs to the technical field of chemical data analysis, and discloses a chemical data analysis system and method for a chemical plant based on industrial big data; the method comprises: real-time acquisition of process parameters of each device in a factory production process, the process parameters comprising temperature, pressure and flow; processing the process parameters of each device and extracting characteristic parameters; judging whether to mark parameters in the characteristic parameters as abnormal parameters; analyzing the flow of the device corresponding to the abnormal parameters and obtaining the corresponding optimal parameters; controlling and adjusting the abnormal parameters of the device according to the optimal parameters; re-acquiring the process parameters of each device and extracting the characteristic parameters, judging again whether to mark the parameters in the characteristic parameters as abnormal parameters, and if abnormal parameters still exist, performing abnormal type analysis and generating early warning instructions; the invention can greatly improve production efficiency and product quality, effectively optimize the production process, and reduce energy consumption and pollution.

Description

Technical Field

[0001] The present invention relates to the technical field of chemical data analysis, and more specifically, to a chemical plant chemical data analysis system and method based on industrial big data. Background Art

[0002] The chemical industry is an important part of the national economy. Traditional data analysis methods in chemical plants usually rely on manual data collection and empirical judgment, which is inefficient and prone to errors. The chemical industry urgently needs to optimize the production process and save energy and reduce emissions through digital transformation and full use of industrial big data technology.

[0003] The Chinese patent with the authorization announcement number CN111984692B discloses a chemical data analysis system based on industrial big data; including a cloud service platform that receives and stores industrial big data information; a data acquisition unit set up in a chemical plant area, and a data statistics module is set on the data acquisition unit; an industrial big data processing unit, and the industrial big data processing unit includes a monitoring and early warning module, a product quality control module, a data analysis module, a modeling module, and a report module; the cloud service platform is linked to the industrial big data processing unit, and the cloud service platform includes a cloud resource unit and a local resource unit, and the local resource unit transmits the chemical (plant area) intranet data to the data acquisition external network interface server of the cloud resource unit, and the cloud resource unit includes a data service API gateway module, a data bus DataHub module, and a big data storage computing service MaxCompute module;

[0004] However, when the above technology is applied to control, it can only give the recommended control parameter setting value range through the established control optimization model, but cannot give the specific control parameters, thus lacking targeted control strategies, resulting in unsatisfactory control effects; it also requires human intervention, cannot respond quickly to process changes, and is difficult to achieve adaptive control; in addition, it is impossible to obtain equipment failure information in a timely manner, thus affecting production efficiency and production quality;

[0005] In view of this, the present invention proposes a chemical plant chemical data analysis system and method based on industrial big data to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: a chemical plant chemical data analysis method based on industrial big data, comprising:

[0007] Real-time collection of process parameters of each device in the factory production process, including temperature, pressure and flow;

[0008] Process the process parameters of each device and extract characteristic parameters;

[0009] Analyze the characteristic parameters of each device to determine whether to mark parameters in the characteristic parameters as abnormal parameters;

[0010] Analyze the traffic of the device corresponding to the abnormal parameters and obtain the corresponding optimal parameters;

[0011] Control and adjust abnormal parameters of equipment according to optimal parameters;

[0012] Re-collect the process parameters of each device and extract the characteristic parameters, and again determine whether to mark the parameters in the characteristic parameters as abnormal parameters. If abnormal parameters still exist, perform abnormal type analysis and generate early warning instructions.

[0013] Further, the process parameters include temperature, pressure and flow rate; the temperature includes inlet temperature and outlet temperature, the pressure includes inlet pressure and outlet pressure, and the flow rate includes inlet flow rate and outlet flow rate;

[0014] The characteristic parameters include temperature characteristic parameters, pressure characteristic parameters and conversion rate; the temperature characteristic parameters include temperature peak value, temperature valley value and temperature difference; the pressure characteristic parameters include pressure peak value, pressure valley value and pressure difference; the conversion rate is the product conversion capacity of the equipment;

[0015] The method for extracting characteristic parameters comprises:

[0016] The inlet temperature of each device is compared with the outlet temperature, and the temperature with a larger value is taken as the temperature peak value of the corresponding device, and the temperature with a smaller value is taken as the temperature valley value of the corresponding device; the inlet pressure of each device is compared with the outlet pressure, and the pressure with a larger value is taken as the pressure peak value of the corresponding device, and the pressure with a smaller value is taken as the pressure valley value of the corresponding device; the difference between the outlet temperature of each device and the corresponding inlet temperature is taken as the temperature difference of the corresponding device; the difference between the outlet pressure of each device and the corresponding inlet pressure is taken as the pressure difference of the corresponding device; the simulation model corresponding to each device is constructed by using modeling and simulation software, and the simulation model is a mathematical and physical model for numerical description and calculation of the equipment operation process; the inlet flow of each device is input into the corresponding simulation model, and the corresponding outlet flow is predicted and marked as the predicted flow; the predicted flow is divided by the outlet flow collected in real time by the corresponding device to obtain the conversion rate of each device.

[0017] Furthermore, the method for determining whether to mark a parameter in the characteristic parameter as an abnormal parameter includes:

[0018] Set a different number for each device, and mark the number of each device as the device number; take the characteristic parameters and device number of a device as a set of analysis data, input each set of analysis data into the trained abnormality judgment model, output the corresponding judgment label, and obtain the corresponding judgment result according to the judgment label, and the judgment result includes temperature abnormality, pressure abnormality, normal and abnormal;

[0019] The training process of the anomaly judgment model includes:

[0020] Pre-set corresponding judgment results for a group of analysis data, where a is an integer greater than 1, and set different digital labels for different judgment results; mark the digital labels of the judgment results as judgment labels, and convert the analysis data and the corresponding judgment labels into a corresponding set of feature vectors; use each set of feature vectors as input to an abnormality judgment model, and the abnormality judgment model uses a set of predicted judgment labels corresponding to each group of analysis data as output, and uses the actual judgment labels corresponding to each group of analysis data as prediction targets, and the actual judgment labels are pre-collected digital labels of the judgment results corresponding to the analysis data; minimize the sum of the prediction errors of all analysis data as a training target; train the abnormality judgment model until the sum of the prediction errors reaches convergence and the training is stopped; the abnormality judgment model is a deep neural network model;

[0021] If the judgment result is that the temperature is abnormal, the temperature characteristic parameter is marked as an abnormal parameter; if the judgment result is that the pressure is abnormal, the pressure characteristic parameter is marked as a pressure abnormality; if the judgment result is that all are abnormal, both the temperature characteristic parameter and the pressure characteristic parameter are marked as abnormal; if the judgment result is that all are normal, no characteristic parameter is marked;

[0022] A corresponding conversion rate threshold is preset for each device, and the conversion rate of each device is compared with the corresponding conversion rate threshold; if the conversion rate is greater than or equal to the conversion rate threshold, the conversion rate is not marked as an abnormal parameter; if the conversion rate is less than the conversion rate threshold, the conversion rate is marked as an abnormal parameter.

[0023] Further, the optimal parameters include optimal inlet temperature, optimal outlet temperature, optimal inlet pressure and optimal outlet pressure;

[0024] The steps of obtaining the corresponding optimal parameters include:

[0025] Step 1: Preset the corresponding parameter range for each device, including the inlet temperature range, outlet temperature range, inlet pressure range and outlet pressure range;

[0026] Step 2: Randomly select a value from each range within the parameter range of each device and combine them into a set of selection parameters corresponding to each device. Set a different digital label for each different set of combined selection parameters and mark them as selection labels. According to the obtained selection labels, construct a selection label set corresponding to each device.

[0027] Step 3: Mark the device corresponding to the abnormal parameter as an abnormal device, unmark the device whose conversion rate is marked as an abnormal parameter among the abnormal devices, randomly select an abnormal device that is not marked as a selected device, and mark it as a selected device. According to the selected label set of the selected device, randomly generate an initial antibody group X(0) = {X1, X2, X3, ... X q}, X q is the qth antibody; preset iteration threshold;

[0028] Step 4: Determine the fitness function;

[0029] Step 5: Calculate the fitness of each antibody in the initial antibody population X(0), select r antibodies with the largest fitness, and construct a memory population X m (t);

[0030] Step 6: Memory population X m Each antibody in (t) is cloned according to the calculated clone size to generate a clone antibody population in is the qth cloned antibody in the clonal antibody group;

[0031] Step 7: Cloning Antibody Group X c Each antibody in (t) is subjected to cloud adaptive mutation using a cloud adaptive mutation operator to generate a mutant antibody group X d (t);

[0032] Step 8: Cloning Antibodies from Population X c (t) randomly selects w antibodies as parent antibodies for antibody recombination to generate offspring antibodies S c , until the cloned antibody group X c When the number of unselected antibodies in (t) is less than w, antibody recombination is stopped and a recombinant antibody group X is generated. r (t);

[0033] Step 9: Merge variant antibody group X d (t) and recombinant antibody group X r (t), generate a combined antibody population X L (t);

[0034] Step 10: Randomly generate n new antibodies and merge antibody group X LThe n antibodies with the smallest fitness in (t) are replaced by n randomly generated new antibodies;

[0035] Step 11: If the number of cycles t is less than the iteration threshold, set t = t + 1 and select the merged antibody group X L The r antibodies with the largest fitness in (t) form a new memory population X m (t+1), loop steps 5 to 10; if the number of iterations t is equal to the iteration threshold, the loop ends and the combined antibody group X is obtained L (t) The selection tag corresponding to the antibody with the largest fitness, obtain the selection parameters corresponding to the screening device according to the selection tag, and use them as the optimal parameters. The screening device is the selected device in step 3;

[0036] Step 12: Repeat steps 3 to 11 until all abnormal devices are marked as selected devices and the cycle ends.

[0037] Furthermore, in step 3, the antibodies in the initial antibody group X(0) correspond one-to-one to the selection tags in the selection tag set, and the range of the antibodies is the selection tag set;

[0038] In step 4, the fitness function is expressed as:

[0039] In the formula, f i is the fitness of the ith antibody, NH i is the energy consumption value of the ith antibody, WR i is the contamination degree of the i-th antibody, ω1 and ω2 are preset proportional coefficients, i∈[1,q];

[0040] The method for obtaining the energy consumption value and the pollution degree is as follows: if the conversion rate of the screening device is marked as an abnormal parameter, the predicted flow rate of the screening device is used as the outlet flow rate; if the conversion rate of the screening device is not marked as an abnormal parameter, the outlet flow rate of the screening device is still the outlet flow rate collected in real time; according to the selection label corresponding to the antibody, the corresponding selection parameters are obtained, and the selection parameters, device number, inlet flow rate and outlet flow rate of the screening device are used as test data; the test data are respectively input into the trained energy consumption prediction model and the pollution prediction model to predict the corresponding energy consumption value and pollution degree;

[0041] The training process of the energy consumption prediction model includes:

[0042] Collect energy consumption values ​​corresponding to c groups of test data in advance, where c is an integer greater than 1, and convert the test data and the corresponding energy consumption values ​​into a corresponding set of feature vectors;

[0043] Each group of feature vectors is used as the input of the energy consumption prediction model. The energy consumption prediction model uses a group of predicted energy consumption values ​​corresponding to each group of test data as output, and uses the actual energy consumption values ​​corresponding to each group of test data as the prediction target. The actual energy consumption values ​​are the energy consumption values ​​corresponding to the test data collected in advance; the prediction results are evaluated for model accuracy using the mean absolute percentage error (MAPE). When the calculated MAPE is less than the preset MAPE, the energy consumption prediction model training is completed; an energy consumption prediction model is generated that predicts energy consumption values ​​based on the test data; the energy consumption prediction model is a deep belief network model;

[0044] The training process of the pollution prediction model is consistent with that of the energy consumption prediction model, and it is also a deep belief network model.

[0045] Furthermore, in step 6, the method for calculating the clone scale includes:

[0046]

[0047] Φ i =min(exp(||X i -X j ||));

[0048] In the formula, R i is the clone size of the i-th antibody, Int is the upward rounding function, q is the size of the antibody group, Φ i is the affinity of the ith antibody, min is the minimum function, exp is the natural exponential function, ||X i -X j || is the Euclidean distance between the i-th antibody and the j-th antibody, i≠j, j∈[1,q];

[0049] In step 8, the progeny antibody S c The calculation methods include:

[0050]

[0051] In the formula, X′ w is the w-th parent antibody, s w is the w-th scaling factor, which is a randomly generated real number and not all w scaling factors are 0.

[0052] Furthermore, the abnormal parameters of the corresponding abnormal equipment are controlled and adjusted according to the obtained optimal parameters; for the equipment whose conversion rate is marked as an abnormal parameter, the outlet flow control is adjusted to the corresponding predicted flow.

[0053] Furthermore, the method for performing abnormal type analysis includes:

[0054] If the abnormal parameter is the temperature peak, the abnormal type of the equipment is judged to be a first-level temperature abnormality;

[0055] If the abnormal parameter is a temperature valley value, the abnormal type of the device is judged to be a secondary temperature abnormality;

[0056] If the abnormal parameter is the temperature difference, the abnormal type of the equipment is judged to be a level 3 temperature abnormality;

[0057] If the abnormal parameter is the pressure peak, the abnormal type of the equipment is judged to be a primary pressure abnormality;

[0058] If the abnormal parameter is a pressure valley value, the abnormal type of the equipment is judged to be a secondary pressure abnormality;

[0059] If the abnormal parameter is the pressure difference, the abnormal type of the equipment is judged to be a third-level pressure abnormality;

[0060] Two adjacent devices in the factory production process are regarded as a group of device sets. In each group of device sets, the device connected to the other device at the inlet is marked as a downstream device, and the device connected to the other device at the outlet is marked as an upstream device; the outlet flow of the upstream device in each group of device sets is subtracted from the inlet flow of the downstream device to obtain the flow difference in each group of device sets; a corresponding difference threshold is preset for each group of device sets, and each flow difference is compared with the corresponding difference threshold; if the flow difference is greater than or equal to the difference threshold, the corresponding device set is marked as an abnormal set; if the flow difference is less than the flow threshold, the corresponding device set is not marked as an abnormal set;

[0061] If the device set is marked as an abnormal set, it is determined that the transmission device between two devices in the device set is abnormal.

[0062] Further, the optimal inlet temperature and the optimal inlet pressure in the optimal parameters corresponding to each device are adjusted;

[0063] Connect the simulation models corresponding to each device to obtain the simulation model corresponding to the factory production line, that is, connect and match the input ports and output ports between the simulation models of each device; input the optimal inlet temperature and optimal inlet pressure in the optimal parameters of each device, the outlet temperature and outlet pressure collected in real time, and the flow rate after control adjustment into the corresponding simulation model; adjust the outlet temperature control of each simulation model to the optimal outlet temperature in the optimal parameters of the corresponding device, adjust the outlet pressure control of each simulation model to the optimal outlet pressure in the optimal parameters of the corresponding device, and collect the inlet temperature and inlet pressure of each simulation model; subtract the collected inlet temperature of each simulation model from the corresponding The best inlet temperature in the optimal parameters of the equipment is used to obtain the inlet temperature difference corresponding to each equipment; the inlet temperature of each simulation model collected is subtracted from the best inlet temperature in the optimal parameters of the corresponding equipment to obtain the inlet temperature difference corresponding to each equipment; the best inlet temperature in the optimal parameters corresponding to each equipment is subtracted from the corresponding inlet temperature difference to obtain the inlet regulating temperature; the best inlet pressure in the optimal parameters corresponding to each equipment is subtracted from the corresponding inlet pressure difference to obtain the inlet regulating pressure; the best inlet temperature in the optimal parameters corresponding to each equipment is adjusted to the corresponding inlet regulating temperature, and the best inlet pressure in the optimal parameters corresponding to each equipment is adjusted to the corresponding inlet regulating pressure.

[0064] A chemical plant chemical data analysis system based on industrial big data, implementing the chemical plant chemical data analysis method based on industrial big data, comprising:

[0065] Parameter collection module, used to collect the process parameters of each equipment in the factory production process in real time, including temperature, pressure and flow;

[0066] Parameter processing module, used to process the process parameters of each device and extract characteristic parameters;

[0067] The abnormality judgment module is used to analyze the characteristic parameters of each device and determine whether to mark the parameters in the characteristic parameters as abnormal parameters;

[0068] The parameter analysis module is used to analyze the traffic of the device corresponding to the abnormal parameters and obtain the corresponding optimal parameters;

[0069] Parameter control module, used to control and adjust abnormal parameters of equipment according to optimal parameters;

[0070] The type analysis module is used to re-collect the process parameters of each device and extract the characteristic parameters, and again determine whether to mark the parameters in the characteristic parameters as abnormal parameters. If abnormal parameters still exist, perform abnormal type analysis and generate early warning instructions.

[0071] The technical effects and advantages of the chemical plant chemical data analysis system and method based on industrial big data of the present invention are as follows:

[0072] 1. Through real-time collection of process parameters, deep learning is used to intelligently analyze equipment anomalies, and clone selection algorithms and machine learning models are combined to obtain the best parameters and automatically adjust the optimization parameters, as well as real-time analysis of anomaly types; it realizes the automatic monitoring and intelligent optimization of chemical data in the chemical production process, can accurately obtain the best parameters for each equipment, and promptly know the failure of equipment and transportation devices, greatly improving production efficiency and product quality; and effectively optimize the production process, reduce energy consumption and pollution; at the same time, it meets the needs of real-time monitoring and efficient optimization.

[0073] 2. By considering the temperature coupling effect and the pressure coupling effect, the simulation model of each device is connected into a system-level simulation model, the difference between the optimal parameters and the simulation results is analyzed, the inlet adjustment temperature and the inlet adjustment pressure are calculated, and then the original optimal parameters are adjusted individually; in order to eliminate the influence of parameter coupling, each device can truly achieve the optimal process parameters, and can achieve parameter optimization more efficiently and accurately, thereby improving operating efficiency and stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0074] Figure 1 This is a schematic diagram of a chemical plant chemical data analysis system based on industrial big data according to Example 1 of the present invention;

[0075] Figure 2 This is a schematic diagram of a chemical data analysis system for a chemical plant based on industrial big data according to Example 2 of the present invention;

[0076] Figure 3 This is a flow chart of a method for analyzing chemical data of a chemical plant based on industrial big data according to Embodiment 3 of the present invention;

[0077] Figure 4 This is a schematic diagram of an electronic device according to Embodiment 4 of the present invention;

[0078] Figure 5 This is a schematic diagram of a storage medium according to Embodiment 5 of the present invention. DETAILED DESCRIPTION

[0079] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0080] Example 1

[0081] See also Figure 1 As shown, this embodiment provides a chemical plant chemical data analysis system based on industrial big data, including a parameter acquisition module, a parameter processing module, an abnormality judgment module, a parameter analysis module, a parameter control module and a type analysis module; each module is connected by wired and / or wireless means to realize data transmission between modules.

[0082] The parameter acquisition module is used to collect the process parameters of each equipment in the factory production process in real time.

[0083] The factory is a chemical factory, and the equipment is production equipment; the process parameters include temperature, pressure and flow; the temperature includes inlet temperature and outlet temperature, the pressure includes inlet pressure and outlet pressure, and the flow includes inlet flow and outlet flow; the inlet temperature, inlet pressure and inlet flow are the temperature, pressure and raw material feed flow at the inlet of the equipment, and the outlet temperature, outlet pressure and outlet flow are the temperature, pressure and product discharge flow at the outlet of the equipment; the inlet temperature and outlet temperature are obtained by temperature sensors installed at the inlet and outlet of each equipment, the inlet pressure and outlet pressure are obtained by pressure sensors installed at the inlet and outlet of each equipment, and the inlet flow and outlet flow are obtained by flow sensors installed at the inlet and outlet of each equipment, and flow sensors such as vortex flowmeters, Doppler flowmeters, rubber membrane flowmeters, photoelectric flowmeters, etc.; among them, vortex flowmeters or Doppler flowmeters can be used for liquids, rubber membrane flowmeters can be used for gases, and photoelectric flowmeters can be used for solids.

[0084] The parameter processing module is used to process the process parameters of each device and extract characteristic parameters.

[0085] Characteristic parameters include temperature characteristic parameters, pressure characteristic parameters and conversion rate; temperature characteristic parameters include temperature peak value, temperature valley value and temperature difference; pressure characteristic parameters include pressure peak value, pressure valley value and pressure difference; conversion rate is the product conversion capacity of the equipment.

[0086] Methods for extracting feature parameters include:

[0087] The inlet temperature of each device is compared with the outlet temperature, and the temperature with a larger value is taken as the temperature peak value of the corresponding device, and the temperature with a smaller value is taken as the temperature valley value of the corresponding device; the inlet pressure of each device is compared with the outlet pressure, and the pressure with a larger value is taken as the pressure peak value of the corresponding device, and the pressure with a smaller value is taken as the pressure valley value of the corresponding device; the difference between the outlet temperature of each device and the corresponding inlet temperature is taken as the temperature difference of the corresponding device; the difference between the outlet pressure of each device and the corresponding inlet pressure is taken as the pressure difference of the corresponding device; TCAD, CoventorWare, COMSOL, Silvaco and other modeling and simulation software are used to build a simulation model corresponding to each device. The simulation model is a mathematical and physical model that numerically describes and calculates the operation process of the device. The construction process of the simulation model is a prior art and will not be elaborated on here. The inlet flow of each device is input into the corresponding simulation model, and the corresponding outlet flow is predicted and marked as the predicted flow; the predicted flow is divided by the outlet flow collected in real time by the corresponding device to obtain the conversion rate of each device.

[0088] The abnormality judgment module is used to analyze the characteristic parameters of each device and determine whether to mark the parameters in the characteristic parameters as abnormal parameters.

[0089] The method for determining whether to mark a parameter in the characteristic parameter as an abnormal parameter includes:

[0090] Set a different number for each device, and mark the number of each device as the device number. For example, the device number of the distillation tower is 10, the device number of the filter press is 20, the device number of the piston compressor is 30, and so on; use the characteristic parameters and device number of a device as a group of analysis data, input each group of analysis data into the trained abnormality judgment model, output the corresponding judgment label, and obtain the corresponding judgment result according to the judgment label, and the judgment results include temperature abnormality, pressure abnormality, normal and abnormal.

[0091] The specific training process of the anomaly judgment model includes:

[0092] A corresponding judgment result is set for a group of analysis data in advance, where a is an integer greater than 1, and different digital labels are set for different judgment results. For example, the digital label is set to 1 for temperature anomaly, the digital label is set to 2 for pressure anomaly, the digital label is set to 3 for all normal conditions, and the digital label is set to 4 for all abnormalities. The judgment result corresponding to the analysis data is collected by a group of different analysis data by technical personnel in the historical chemical data analysis process, and the technical personnel in this field judge in turn under the conditions of each group of analysis data according to whether there is an abnormality in the parameters in the characteristic parameters based on actual experience and equipment technical specifications, and set corresponding judgment results for a group of different analysis data in turn.

[0093] The digital label of the judgment result is marked as the judgment label, and the analysis data and the corresponding judgment label are converted into a corresponding set of feature vectors; each set of feature vectors is used as the input of the abnormal judgment model, and the abnormal judgment model uses a set of predicted judgment labels corresponding to each set of analysis data as output, and uses the actual judgment label corresponding to each set of analysis data as the prediction target, and the actual judgment label is the digital label of the judgment result corresponding to the analysis data collected in advance; the training goal is to minimize the sum of the prediction errors of all analysis data; wherein the calculation formula of the prediction error is Z K =(α K -μ K ) 2 , where Z K is the prediction error, K is the group number of the eigenvector corresponding to the analysis data, α K is the predicted judgment label corresponding to the Kth group of analysis data, μ K is the actual judgment label corresponding to the Kth group of analysis data; the abnormal judgment model is trained until the sum of the prediction errors reaches convergence and the training is stopped; the abnormal judgment model is specifically a deep neural network model.

[0094] If the judgment result is temperature abnormality, the temperature characteristic parameter is marked as an abnormal parameter; if the judgment result is pressure abnormality, the pressure characteristic parameter is marked as pressure abnormality; if the judgment result is all abnormal, both the temperature characteristic parameter and the pressure characteristic parameter are marked as abnormal; if the judgment result is all normal, the characteristic parameter is not marked.

[0095] A corresponding conversion rate threshold is preset for each device, and the conversion rate of each device is compared with the corresponding conversion rate threshold; if the conversion rate is greater than or equal to the conversion rate threshold, the conversion rate is not marked as an abnormal parameter; if the conversion rate is less than the conversion rate threshold, the conversion rate is marked as an abnormal parameter; the conversion rate threshold of each device is pre-set by technical personnel in this field according to the technical specifications of each device.

[0096] The parameter analysis module is used to analyze the traffic of the device corresponding to the abnormal parameters and obtain the corresponding optimal parameters.

[0097] The optimal parameters include optimal inlet temperature, optimal outlet temperature, optimal inlet pressure and optimal outlet pressure.

[0098] The steps of obtaining the corresponding optimal parameters include:

[0099] Step 1: A corresponding parameter range is preset for each device, the parameter range includes an inlet temperature range, an outlet temperature range, an inlet pressure range and an outlet pressure range, and the parameter range is preset by a technician in this field according to the technical specifications of the device;

[0100] Step 2: Randomly select a value from each range within the parameter range of each device and combine them into a set of selection parameters corresponding to each device. Set a different digital label for each different set of combined selection parameters and mark them as selection labels. According to the obtained selection labels, construct a selection label set corresponding to each device.

[0101] Step 3: Mark the device corresponding to the abnormal parameter as an abnormal device, unmark the device whose conversion rate is marked as an abnormal parameter among the abnormal devices, randomly select an abnormal device that is not marked as a selected device, and mark it as a selected device. According to the selected label set of the selected device, randomly generate an initial antibody group X(0) = {X1, X2, X3, ... X q}, X q is the qth antibody; a preset iteration threshold is preset according to the algorithm accuracy;

[0102] Step 4: Determine the fitness function;

[0103] Step 5: Calculate the fitness of each antibody in the initial antibody population X(0), select r antibodies with the largest fitness, and construct a memory population X m (t);

[0104] Step 6: Memory population X m Each antibody in (t) is cloned according to the calculated clone size to generate a clone antibody population in is the qth cloned antibody in the clonal antibody group;

[0105] Step 7: Cloning Antibody Group X c Each antibody in (t) is subjected to cloud adaptive mutation using a cloud adaptive mutation operator to generate a mutant antibody group X d (t); The cloud adaptive mutation operator is an existing technology and will not be described in detail here;

[0106] Step 8: Cloning Antibodies from Population X c (t) randomly selects w antibodies as parent antibodies for antibody recombination to generate offspring antibodies S c , until the cloned antibody group X c When the number of unselected antibodies in (t) is less than w, antibody recombination is stopped and a recombinant antibody group X is generated. r (t); In this embodiment, w is preferably 3; recombinant antibody group X r (t) includes multiple progeny antibodies S c ;

[0107] Step 9: Merge variant antibody group X d (t) and recombinant antibody group X r(t), generate a combined antibody population X L (t);

[0108] Step 10: Randomly generate n new antibodies and merge antibody group X L The n antibodies with the smallest fitness in (t) are replaced by n randomly generated new antibodies;

[0109] Step 11: If the number of cycles t is less than the iteration threshold, set t = t + 1 and select the merged antibody group X L The r antibodies with the largest fitness in (t) form a new memory population X m (t+1), loop steps 5 to 10; if the number of iterations t is equal to the iteration threshold, the loop ends and the combined antibody group X is obtained L (t) The selection tag corresponding to the antibody with the largest fitness, the selection parameter corresponding to the screening device is obtained according to the selection tag, and is used as the optimal parameter, and the screening device is the selected device in step 3; wherein t=t+1 is a computer language, that is, the value of the iteration number t is added by 1 and then assigned to the iteration number t;

[0110] Step 12: Repeat steps 3 to 11 until all abnormal devices are marked as selected devices and the cycle ends.

[0111] In the above step 3, the antibodies in the initial antibody group X(0) correspond one-to-one to the selection labels in the selection label set, and the range of the antibodies is the selection label set.

[0112] In step 4 above, the fitness function is expressed as:

[0113] In the formula, f i is the fitness of the ith antibody, NH i is the energy consumption value of the ith antibody, WR i is the contamination degree of the i-th antibody, ω1 and ω2 are preset proportional coefficients, i∈[1,q];

[0114] It should be noted that the specific value of the proportionality coefficient in the formula can be set according to actual conditions. The proportionality coefficient reflects the importance of the energy consumption value and the pollution degree. Technical personnel in this field can preset the corresponding weight coefficient according to the actual importance of the energy consumption value and the pollution degree, so as to optimize energy consumption while reducing environmental pollution; fitness is only an evaluation criterion for screening antibodies, so the calculation process of fitness is a dimensionless calculation.

[0115] The method for obtaining the energy consumption value and pollution degree is as follows: if the conversion rate of the screening device is marked as an abnormal parameter, the predicted flow rate of the screening device is used as the outlet flow rate; if the conversion rate of the screening device is not marked as an abnormal parameter, the outlet flow rate of the screening device is still the outlet flow rate collected in real time; according to the selection label corresponding to the antibody, the corresponding selection parameters are obtained, and the selection parameters, device number, import flow rate and outlet flow rate of the screening device are used as test data; the test data are respectively input into the trained energy consumption prediction model and pollution prediction model to predict the corresponding energy consumption value and pollution degree.

[0116] The specific training process of the energy consumption prediction model includes:

[0117] Energy consumption values ​​corresponding to c groups of test data are collected in advance, and the test data and the corresponding energy consumption values ​​are converted into a corresponding set of feature vectors.

[0118] Each set of feature vectors is used as the input of the energy consumption prediction model. The energy consumption prediction model uses a set of predicted energy consumption values ​​corresponding to each set of test data as output, and uses the actual energy consumption values ​​corresponding to each set of test data as the prediction target. The actual energy consumption value is the energy consumption value corresponding to the test data collected in advance. The mean absolute percentage error MAPE is used to evaluate the model accuracy of the prediction results. When the calculated MAPE is less than the preset MAPE, the energy consumption prediction model training is completed. The calculation formula of MAPE is: Among them, β L is the actual energy consumption value corresponding to the Lth group of test data, δ L is the predicted energy consumption value corresponding to the Lth group of test data, L is the group number of the feature vector corresponding to the test data, H represents the number of predicted energy consumption values, L∈[1,H]; an energy consumption prediction model that predicts energy consumption values ​​according to the test data is generated; wherein the energy consumption prediction model is a deep belief network model, and the preset MAPE is pre-set by technicians in this field according to the accuracy required by the energy consumption prediction model.

[0119] It should be noted that the energy consumption values ​​corresponding to the test data are obtained by technical personnel in this field during the historical chemical data analysis process. They collect c groups of different test data, perform multiple tests and collect corresponding energy consumption values ​​under the conditions of each group of test data, and take the average of the multiple collected energy consumption values ​​as the energy consumption value of the corresponding test data. In this way, energy consumption values ​​corresponding to c groups of test data are collected.

[0120] The specific training process of the pollution prediction model is consistent with that of the energy consumption prediction model, and it is also a deep belief network model.

[0121] In the above step 6, the calculation method of the clone size includes:

[0122]

[0123] Φ i =min(exp(||X i -X j ||));

[0124] In the formula, R i is the clone size of the i-th antibody, Int is the upward rounding function, q is the size of the antibody group, Φ i is the affinity of the ith antibody, min is the minimum function, exp is the natural exponential function, ||X i -X j || is the Euclidean distance between the i-th antibody and the j-th antibody, i≠j, j∈[1,q].

[0125] It should be noted that when calculating the Euclidean distance between the ith antibody and the jth antibody, normalization is performed so that 0≤||X i -X j ||≤1.

[0126] In step 8 above, the progeny antibody S c The calculation methods include:

[0127]

[0128] In the formula, X′ w is the w-th parent antibody, s w is the w-th scaling factor, which is a randomly generated real number and not all w scaling factors are 0.

[0129] Parameter control module, used to control and adjust abnormal parameters of equipment according to optimal parameters;

[0130] According to the obtained optimal parameters, the abnormal parameters of the corresponding abnormal equipment are controlled and adjusted respectively; for the equipment whose conversion rate is marked as abnormal parameters, the outlet flow control is adjusted to the corresponding predicted flow.

[0131] The type analysis module is used to re-collect the process parameters of each device and extract the characteristic parameters, and again determine whether to mark the parameters in the characteristic parameters as abnormal parameters. If abnormal parameters still exist, perform abnormal type analysis and generate early warning instructions.

[0132] Methods for performing anomaly type analysis include:

[0133] If the abnormal parameter is the temperature peak, the abnormal type of the equipment is judged to be a level 1 temperature abnormality, which means that the temperature peak is too high, and there may be abnormal types such as heat accumulation, material aging, or cooling equipment failure;

[0134] If the abnormal parameter is the temperature valley value, the abnormal type of the equipment is judged to be a secondary temperature abnormality; this means that the temperature valley value is too low, and there may be abnormal types such as heating equipment failure or heat loss;

[0135] If the abnormal parameter is the temperature difference, the abnormal type of the equipment is judged to be a level 3 temperature abnormality, which means that the temperature difference is too large, and there may be abnormal types such as reduced heat transfer efficiency or heat leakage;

[0136] If the abnormal parameter is the pressure peak, the abnormal type of the equipment is judged to be a primary pressure abnormality; this means that the pressure peak is too high, and there may be abnormal types such as pipeline blockage, valve failure or compressor failure;

[0137] If the abnormal parameter is the pressure valley value, the abnormal type of the equipment is judged to be a secondary pressure abnormality; this means that the pressure valley value is too low, and there may be abnormal types such as pump or compressor failure;

[0138] If the abnormal parameter is the pressure difference, the abnormal type of the equipment is judged to be a third-level pressure abnormality; this means that the pressure difference is too large, and there may be abnormal types such as increased pipeline resistance or equipment wear.

[0139] Two adjacent devices in the factory production process are regarded as a group of device sets. In each device set, the device connected to the other device at the inlet is marked as a downstream device, and the device connected to the other device at the outlet is marked as an upstream device. For example, there are three devices in the factory production process, device 1 is connected to device 2, and device 2 is connected to device 3. Therefore, device 1 and device 2 are regarded as a group of device sets, and device 2 and device 3 are also regarded as a group of device sets. In the device set of device 1 and device 2, the outlet of device 1 is connected to the inlet of device 2. Therefore, device 1 is an upstream device and device 2 is a downstream device. The upstream device in each device set is marked as the downstream device. The outlet flow of the equipment is subtracted from the inlet flow of the downstream equipment to obtain the flow difference in each equipment set; a corresponding difference threshold is preset for each equipment set, and each flow difference is compared with the corresponding difference threshold; if the flow difference is greater than or equal to the difference threshold, the corresponding equipment set is marked as an abnormal set; if the flow difference is less than the flow threshold, the corresponding equipment set is not marked as an abnormal set; the difference threshold of each equipment set is determined by technicians in this field in the historical chemical data analysis process, who collect the flow difference between the two equipments in each equipment set when the transmission device is abnormal for many times, and take the average of multiple flow differences corresponding to each equipment set as the difference threshold of each equipment set.

[0140] If the device set is marked as an abnormal set, it is judged that the transmission device between the two devices in the device set is abnormal; this means that the flow difference is too large, the outlet flow of the upstream device is much larger than the inlet flow of the downstream device, and there may be pipeline leakage or seepage, delivery pump failure and other abnormal types.

[0141] It should be noted that the early warning instructions are sent by the chemical data analysis system to the mobile devices of relevant operation and maintenance personnel, so that the relevant operation and maintenance personnel can inspect and repair abnormal equipment or abnormal transportation equipment in a timely manner.

[0142] This embodiment collects process parameters in real time, uses deep learning to intelligently analyze equipment anomalies, combines clone selection algorithms and machine learning models to obtain optimal parameters and automatically adjust optimization parameters, as well as real-time analysis of anomaly types; realizes automated monitoring and intelligent optimization of chemical data in the chemical production process, can accurately obtain the optimal parameters of each device, and promptly learn about equipment and transportation device failures, greatly improving production efficiency and product quality; and effectively optimizes the production process, reduces energy consumption and pollution; while meeting the needs for real-time monitoring and efficient optimization.

[0143] Example 2

[0144] See also Figure 2 As shown, this embodiment further improves the design on the basis of embodiment 1. In embodiment 1, the inlet temperature and inlet pressure of each device are controlled and adjusted according to the obtained optimal parameters. However, in actual situations, due to the existence of temperature coupling effect and pressure coupling effect, that is, in a set of equipment, the outlet temperature and outlet pressure of the upstream equipment will affect the inlet temperature and inlet pressure of the downstream equipment. If the inlet temperature and inlet pressure of each device are directly controlled and adjusted according to the obtained optimal parameters, the adjusted inlet temperature and inlet pressure still cannot reach the optimal parameters. Therefore, this embodiment provides a chemical plant chemical data analysis system based on industrial big data, which also includes a parameter adjustment module;

[0145] The parameter adjustment module is used to adjust the optimal inlet temperature and the optimal inlet pressure in the optimal parameters corresponding to each device.

[0146] Connect the simulation models corresponding to each device to obtain the simulation model corresponding to the factory production line, that is, dock and match the input port and output port between the simulation models of each device, and connect the simulation models as the existing technology, which will not be elaborated here; input the optimal inlet temperature and optimal inlet pressure in the optimal parameters of each device, the outlet temperature and outlet pressure collected in real time, and the flow rate after control adjustment into the corresponding simulation model; adjust the outlet temperature control of each simulation model to the optimal outlet temperature in the optimal parameters of the corresponding device, adjust the outlet pressure control of each simulation model to the optimal outlet pressure in the optimal parameters of the corresponding device, and collect the inlet temperature and inlet pressure of each simulation model; input the collected flow rate of each simulation model into the corresponding simulation model; The inlet temperature of the simulated model is subtracted from the best inlet temperature in the optimal parameters of the corresponding equipment to obtain the inlet temperature difference corresponding to each equipment; the inlet temperature of each simulated model collected is subtracted from the best inlet temperature in the optimal parameters of the corresponding equipment to obtain the inlet temperature difference corresponding to each equipment; the best inlet temperature in the optimal parameters corresponding to each equipment is subtracted from the corresponding inlet temperature difference to obtain the inlet regulating temperature; the best inlet pressure in the optimal parameters corresponding to each equipment is subtracted from the corresponding inlet pressure difference to obtain the inlet regulating pressure; the best inlet temperature in the optimal parameters corresponding to each equipment is adjusted to the corresponding inlet regulating temperature, and the best inlet pressure in the optimal parameters corresponding to each equipment is adjusted to the corresponding inlet regulating pressure.

[0147] This embodiment connects the simulation model of each device into a system-level simulation model by considering the temperature coupling effect and the pressure coupling effect, analyzes the difference between the optimal parameters and the simulation results, calculates the inlet regulating temperature and the inlet regulating pressure, and then makes personalized adjustments to the original optimal parameters; so as to eliminate the influence of parameter coupling, enable each device to truly achieve the optimal process parameters, and realize parameter optimization more efficiently and accurately, thereby improving operating efficiency and stability.

[0148] Example 3

[0149] See also Figure 3 As shown, the part not described in detail in this embodiment is described in Examples 1 and 2, and a method for analyzing chemical data of a chemical plant based on industrial big data is provided, the method comprising:

[0150] Real-time collection of process parameters of each device in the factory production process, including temperature, pressure and flow;

[0151] Process the process parameters of each device and extract characteristic parameters;

[0152] Analyze the characteristic parameters of each device to determine whether to mark parameters in the characteristic parameters as abnormal parameters;

[0153] Analyze the traffic of the device corresponding to the abnormal parameters and obtain the corresponding optimal parameters;

[0154] Control and adjust abnormal parameters of equipment according to optimal parameters;

[0155] Re-collect the process parameters of each device and extract the characteristic parameters, and again determine whether to mark the parameters in the characteristic parameters as abnormal parameters. If abnormal parameters still exist, perform abnormal type analysis and generate early warning instructions.

[0156] Furthermore, the process parameters include temperature, pressure and flow; the temperature includes inlet temperature and outlet temperature, the pressure includes inlet pressure and outlet pressure, and the flow includes inlet flow and outlet flow.

[0157] The characteristic parameters include temperature characteristic parameters, pressure characteristic parameters and conversion rate; the temperature characteristic parameters include temperature peak value, temperature valley value and temperature difference; the pressure characteristic parameters include pressure peak value, pressure valley value and pressure difference; the conversion rate is the product conversion capacity of the equipment;

[0158] The method for extracting characteristic parameters comprises:

[0159] The inlet temperature of each device is compared with the outlet temperature, and the temperature with a larger value is taken as the temperature peak value of the corresponding device, and the temperature with a smaller value is taken as the temperature valley value of the corresponding device; the inlet pressure of each device is compared with the outlet pressure, and the pressure with a larger value is taken as the pressure peak value of the corresponding device, and the pressure with a smaller value is taken as the pressure valley value of the corresponding device; the difference between the outlet temperature of each device and the corresponding inlet temperature is taken as the temperature difference of the corresponding device; the difference between the outlet pressure of each device and the corresponding inlet pressure is taken as the pressure difference of the corresponding device; the simulation model corresponding to each device is constructed by using modeling and simulation software, and the simulation model is a mathematical and physical model for numerical description and calculation of the equipment operation process; the inlet flow of each device is input into the corresponding simulation model, and the corresponding outlet flow is predicted and marked as the predicted flow; the predicted flow is divided by the outlet flow collected in real time by the corresponding device to obtain the conversion rate of each device.

[0160] Furthermore, the method for determining whether to mark a parameter in the characteristic parameter as an abnormal parameter includes:

[0161] Set a different number for each device, and mark the number of each device as the device number; take the characteristic parameters and device number of a device as a set of analysis data, input each set of analysis data into the trained abnormality judgment model, output the corresponding judgment label, and obtain the corresponding judgment result according to the judgment label, and the judgment result includes temperature abnormality, pressure abnormality, normal and abnormal;

[0162] The training process of the anomaly judgment model includes:

[0163] Pre-set corresponding judgment results for a group of analysis data, where a is an integer greater than 1, and set different digital labels for different judgment results; mark the digital labels of the judgment results as judgment labels, and convert the analysis data and the corresponding judgment labels into a corresponding set of feature vectors; use each set of feature vectors as input to an abnormality judgment model, and the abnormality judgment model uses a set of predicted judgment labels corresponding to each group of analysis data as output, and uses the actual judgment labels corresponding to each group of analysis data as prediction targets, and the actual judgment labels are pre-collected digital labels of the judgment results corresponding to the analysis data; minimize the sum of the prediction errors of all analysis data as a training target; train the abnormality judgment model until the sum of the prediction errors reaches convergence and the training is stopped; the abnormality judgment model is a deep neural network model;

[0164] If the judgment result is that the temperature is abnormal, the temperature characteristic parameter is marked as an abnormal parameter; if the judgment result is that the pressure is abnormal, the pressure characteristic parameter is marked as a pressure abnormality; if the judgment result is that all are abnormal, both the temperature characteristic parameter and the pressure characteristic parameter are marked as abnormal; if the judgment result is that all are normal, no characteristic parameter is marked;

[0165] A corresponding conversion rate threshold is preset for each device, and the conversion rate of each device is compared with the corresponding conversion rate threshold; if the conversion rate is greater than or equal to the conversion rate threshold, the conversion rate is not marked as an abnormal parameter; if the conversion rate is less than the conversion rate threshold, the conversion rate is marked as an abnormal parameter.

[0166] Further, the optimal parameters include optimal inlet temperature, optimal outlet temperature, optimal inlet pressure and optimal outlet pressure;

[0167] The steps of obtaining the corresponding optimal parameters include:

[0168] Step 1: Preset the corresponding parameter range for each device, including the inlet temperature range, outlet temperature range, inlet pressure range and outlet pressure range;

[0169] Step 2: Randomly select a value from each range within the parameter range of each device and combine them into a set of selection parameters corresponding to each device. Set a different digital label for each different set of combined selection parameters and mark them as selection labels. According to the obtained selection labels, construct a selection label set corresponding to each device.

[0170] Step 3: Mark the device corresponding to the abnormal parameter as an abnormal device, unmark the device whose conversion rate is marked as an abnormal parameter among the abnormal devices, randomly select an abnormal device that is not marked as a selected device, and mark it as a selected device. According to the selected label set of the selected device, randomly generate an initial antibody group X(0) = {X1, X2, X3, ... X q}, X q is the qth antibody; preset iteration threshold;

[0171] Step 4: Determine the fitness function;

[0172] Step 5: Calculate the fitness of each antibody in the initial antibody population X(0), select r antibodies with the largest fitness, and construct a memory population X m (t);

[0173] Step 6: Memory population X m Each antibody in (t) is cloned according to the calculated clone size to generate a clone antibody population in is the qth cloned antibody in the clonal antibody group;

[0174] Step 7: Cloning Antibody Group X c Each antibody in (t) is subjected to cloud adaptive mutation using a cloud adaptive mutation operator to generate a mutant antibody group X d (t);

[0175] Step 8: Cloning Antibodies from Population X c (t) randomly selects w antibodies as parent antibodies for antibody recombination to generate offspring antibodies S c , until the cloned antibody group X c When the number of unselected antibodies in (t) is less than w, antibody recombination is stopped and a recombinant antibody group X is generated. r (t);

[0176] Step 9: Merge variant antibody group X d (t) and recombinant antibody group X r (t), generate a combined antibody population X L (t);

[0177] Step 10: Randomly generate n new antibodies and merge antibody group X L The n antibodies with the smallest fitness in (t) are replaced by n randomly generated new antibodies;

[0178] Step 11: If the number of cycles t is less than the iteration threshold, set t = t + 1 and select the merged antibody group X L The r antibodies with the largest fitness in (t) form a new memory population X m(t+1), loop steps 5 to 10; if the number of iterations t is equal to the iteration threshold, the loop ends and the combined antibody group X is obtained L (t) The selection tag corresponding to the antibody with the largest fitness, obtain the selection parameters corresponding to the screening device according to the selection tag, and use them as the optimal parameters. The screening device is the selected device in step 3;

[0179] Step 12: Repeat steps 3 to 11 until all abnormal devices are marked as selected devices and the cycle ends.

[0180] Furthermore, in step 3, the antibodies in the initial antibody group X(0) correspond one-to-one to the selection tags in the selection tag set, and the range of the antibodies is the selection tag set;

[0181] In step 4, the fitness function is expressed as:

[0182] In the formula, f i is the fitness of the ith antibody, NH i is the energy consumption value of the ith antibody, WR i is the contamination degree of the i-th antibody, ω1 and ω2 are preset proportional coefficients, i∈[1,q];

[0183] The method for obtaining the energy consumption value and the pollution degree is as follows: if the conversion rate of the screening device is marked as an abnormal parameter, the predicted flow rate of the screening device is used as the outlet flow rate; if the conversion rate of the screening device is not marked as an abnormal parameter, the outlet flow rate of the screening device is still the outlet flow rate collected in real time; according to the selection label corresponding to the antibody, the corresponding selection parameters are obtained, and the selection parameters, device number, inlet flow rate and outlet flow rate of the screening device are used as test data; the test data are respectively input into the trained energy consumption prediction model and the pollution prediction model to predict the corresponding energy consumption value and pollution degree;

[0184] The training process of the energy consumption prediction model includes:

[0185] Collect energy consumption values ​​corresponding to c groups of test data in advance, where c is an integer greater than 1, and convert the test data and the corresponding energy consumption values ​​into a corresponding set of feature vectors;

[0186] Each group of feature vectors is used as the input of the energy consumption prediction model. The energy consumption prediction model uses a group of predicted energy consumption values ​​corresponding to each group of test data as output, and uses the actual energy consumption values ​​corresponding to each group of test data as the prediction target. The actual energy consumption values ​​are the energy consumption values ​​corresponding to the test data collected in advance; the prediction results are evaluated for model accuracy using the mean absolute percentage error (MAPE). When the calculated MAPE is less than the preset MAPE, the energy consumption prediction model training is completed; an energy consumption prediction model is generated that predicts energy consumption values ​​based on the test data; the energy consumption prediction model is a deep belief network model;

[0187] The training process of the pollution prediction model is the same as that of the energy consumption prediction model, and it is also a deep belief network model;

[0188] Furthermore, in step 6, the method for calculating the clone scale includes:

[0189]

[0190] Φ i =min(exp(||X i -X j ||));

[0191] In the formula, R i is the clone size of the i-th antibody, Int is the upward rounding function, q is the size of the antibody group, Φ i is the affinity of the ith antibody, min is the minimum function, exp is the natural exponential function, ||X i -X j || is the Euclidean distance between the i-th antibody and the j-th antibody, i≠j, j∈[1,q];

[0192] In step 8, the progeny antibody S c The calculation methods include:

[0193]

[0194] In the formula, X′ w is the w-th parent antibody, s w is the w-th scaling factor, which is a randomly generated real number and not all w scaling factors are 0.

[0195] Furthermore, the abnormal parameters of the corresponding abnormal equipment are controlled and adjusted according to the obtained optimal parameters; for the equipment whose conversion rate is marked as an abnormal parameter, the outlet flow control is adjusted to the corresponding predicted flow.

[0196] Furthermore, the method for performing abnormal type analysis includes:

[0197] If the abnormal parameter is the temperature peak, the abnormal type of the equipment is judged to be a first-level temperature abnormality;

[0198] If the abnormal parameter is a temperature valley value, the abnormal type of the device is judged to be a secondary temperature abnormality;

[0199] If the abnormal parameter is the temperature difference, the abnormal type of the equipment is judged to be a level 3 temperature abnormality;

[0200] If the abnormal parameter is the pressure peak, the abnormal type of the equipment is judged to be a primary pressure abnormality;

[0201] If the abnormal parameter is a pressure valley value, the abnormal type of the equipment is judged to be a secondary pressure abnormality;

[0202] If the abnormal parameter is the pressure difference, the abnormal type of the equipment is judged to be a third-level pressure abnormality;

[0203] Two adjacent devices in the factory production process are regarded as a group of device sets. In each group of device sets, the device connected to the other device at the inlet is marked as a downstream device, and the device connected to the other device at the outlet is marked as an upstream device; the outlet flow of the upstream device in each group of device sets is subtracted from the inlet flow of the downstream device to obtain the flow difference in each group of device sets; a corresponding difference threshold is preset for each group of device sets, and each flow difference is compared with the corresponding difference threshold; if the flow difference is greater than or equal to the difference threshold, the corresponding device set is marked as an abnormal set; if the flow difference is less than the flow threshold, the corresponding device set is not marked as an abnormal set;

[0204] If the device set is marked as an abnormal set, it is determined that the transmission device between two devices in the device set is abnormal.

[0205] Further, the optimal inlet temperature and the optimal inlet pressure in the optimal parameters corresponding to each device are adjusted;

[0206] Connect the simulation models corresponding to each device to obtain the simulation model corresponding to the factory production line, that is, connect and match the input ports and output ports between the simulation models of each device; input the optimal inlet temperature and optimal inlet pressure in the optimal parameters of each device, the outlet temperature and outlet pressure collected in real time, and the flow rate after control adjustment into the corresponding simulation model; adjust the outlet temperature control of each simulation model to the optimal outlet temperature in the optimal parameters of the corresponding device, adjust the outlet pressure control of each simulation model to the optimal outlet pressure in the optimal parameters of the corresponding device, and collect the inlet temperature and inlet pressure of each simulation model; subtract the collected inlet temperature of each simulation model from the corresponding The best inlet temperature in the optimal parameters of the equipment is used to obtain the inlet temperature difference corresponding to each equipment; the inlet temperature of each simulation model collected is subtracted from the best inlet temperature in the optimal parameters of the corresponding equipment to obtain the inlet temperature difference corresponding to each equipment; the best inlet temperature in the optimal parameters corresponding to each equipment is subtracted from the corresponding inlet temperature difference to obtain the inlet regulating temperature; the best inlet pressure in the optimal parameters corresponding to each equipment is subtracted from the corresponding inlet pressure difference to obtain the inlet regulating pressure; the best inlet temperature in the optimal parameters corresponding to each equipment is adjusted to the corresponding inlet regulating temperature, and the best inlet pressure in the optimal parameters corresponding to each equipment is adjusted to the corresponding inlet regulating pressure.

[0207] Example 4

[0208] See also Figure 4 As shown, according to another aspect of the present application, an electronic device 500 is also provided. The electronic device 500 may include one or more processors and one or more memories. The memories store computer readable codes, and when the computer readable codes are executed by one or more processors, a chemical plant chemical data analysis method based on industrial big data as described above may be executed.

[0209] The method or system according to the embodiment of the present application can also be used by Figure 4 The electronic device architecture shown in FIG. Figure 4 As shown, the electronic device 500 may include a bus 501, one or more CPUs 502, ROM 503, RAM 504, a communication port 505 connected to a network, an input / output 506, a hard disk 507, etc. The storage device in the electronic device 500, such as ROM 503 or hard disk 507, may store a chemical plant chemical data analysis method based on industrial big data provided by the present application. Furthermore, the electronic device 500 may also include a user interface 508. Of course, Figure 4 The architecture shown is only exemplary and can be omitted according to actual needs when implementing different devices. Figure 4 One or more components of an electronic device are shown.

[0210] Example 5

[0211] See also Figure 5 As shown, a computer-readable storage medium 600 according to an embodiment of the present application is shown. Computer-readable instructions are stored on the computer-readable storage medium 600. When the computer-readable instructions are executed by the processor, a chemical plant chemical data analysis method based on industrial big data according to an embodiment of the present application described with reference to the above figures can be executed. The storage medium 600 includes, but is not limited to, for example, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and cache memory (cache), etc. Non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0212] In addition, according to the embodiments of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the present application provides a non-transitory machine-readable storage medium, which stores machine-readable instructions, and the machine-readable instructions can be executed by a processor to execute instructions corresponding to the method steps provided in the present application, for example: a chemical plant chemical data analysis method based on industrial big data. When the computer program is executed by a central processing unit (CPU), the above functions defined in the method of the present application are executed.

[0213] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0214] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A chemical plant chemical data analysis method based on industrial big data, characterized in that: include: Real-time collection of process parameters of each device in the factory production process, including temperature, pressure and flow; Process the process parameters of each device and extract characteristic parameters; Analyze the characteristic parameters of each device to determine whether to mark parameters in the characteristic parameters as abnormal parameters; Analyze the traffic of the device corresponding to the abnormal parameters and obtain the corresponding optimal parameters; The optimal parameters include optimal inlet temperature, optimal outlet temperature, optimal inlet pressure and optimal outlet pressure; The steps of obtaining the corresponding optimal parameters include: Step 1: Preset the corresponding parameter range for each device, including the inlet temperature range, outlet temperature range, inlet pressure range and outlet pressure range; Step 2: Randomly select a value from each range within the parameter range of each device and combine them into a set of selection parameters corresponding to each device. Set a different digital label for each different set of combined selection parameters and mark them as selection labels. According to the obtained selection labels, construct a selection label set corresponding to each device. Step 3: Mark the device corresponding to the abnormal parameter as an abnormal device, unmark the device whose conversion rate is marked as an abnormal parameter among the abnormal devices, randomly select an abnormal device that is not marked as a selected device, and mark it as a selected device. According to the selected label set of the selected device, randomly generate an initial antibody group X(0) = {X1, X2, X3, ... X q }, X q is the qth antibody; preset iteration threshold; Step 4: Determine the fitness function; Step 5: Calculate the fitness of each antibody in the initial antibody population X(0), select r antibodies with the largest fitness, and construct a memory population X m (t); Step 6: Memory population X m Each antibody in (t) is cloned according to the calculated clone size to generate a clone antibody population in is the qth cloned antibody in the clonal antibody group; Step 7: Cloning Antibody Group X c Each antibody in (t) is subjected to cloud adaptive mutation using a cloud adaptive mutation operator to generate a mutant antibody group X d (t); Step 8: Cloning Antibodies from Population X c (t) randomly selects w antibodies as parent antibodies for antibody recombination to generate offspring antibodies S c , until the cloned antibody group X c When the number of unselected antibodies in (t) is less than w, antibody recombination is stopped and a recombinant antibody group X is generated. r (t); Step 9: Merge variant antibody group X d (t) and recombinant antibody group X r (t), generate a combined antibody population X L (t); Step 10: Randomly generate n new antibodies and merge antibody group X L The n antibodies with the smallest fitness in (t) are replaced by n randomly generated new antibodies; Step 11: If the number of cycles t is less than the iteration threshold, set t = t + 1 and select the merged antibody group X L The r antibodies with the largest fitness in (t) form a new memory population X m (t+1), loop steps 5 to 10; if the number of iterations t is equal to the iteration threshold, the loop ends and the combined antibody group X is obtained. L (t) The selection tag corresponding to the antibody with the largest fitness, the selection parameters corresponding to the screening device are obtained according to the selection tag and used as the optimal parameters. The screening device is the selected device in step 3; Step 12: loop step 3 to step 11 until all abnormal devices are marked as selected devices, and the loop ends; In step 3, the antibodies in the initial antibody group X(0) correspond one-to-one to the selection tags in the selection tag set, and the range of the antibodies is the selection tag set; In step 4, the fitness function is expressed as: In the formula, f i is the fitness of the ith antibody, NH i is the energy consumption value of the ith antibody, WR i is the contamination degree of the i-th antibody, ω1 and ω2 are preset proportional coefficients, i∈[1,q]; The method for obtaining the energy consumption value and the pollution degree is as follows: if the conversion rate of the screening device is marked as an abnormal parameter, the predicted flow rate of the screening device is used as the outlet flow rate; if the conversion rate of the screening device is not marked as an abnormal parameter, the outlet flow rate of the screening device is still the outlet flow rate collected in real time; according to the selection label corresponding to the antibody, the corresponding selection parameters are obtained, and the selection parameters, device number, inlet flow rate and outlet flow rate of the screening device are used as test data; the test data are respectively input into the trained energy consumption prediction model and the pollution prediction model to predict the corresponding energy consumption value and pollution degree; The training process of the energy consumption prediction model includes: Collect energy consumption values ​​corresponding to c groups of test data in advance, where c is an integer greater than 1, and convert the test data and the corresponding energy consumption values ​​into a corresponding set of feature vectors; Each group of feature vectors is used as the input of the energy consumption prediction model. The energy consumption prediction model uses a group of predicted energy consumption values ​​corresponding to each group of test data as output, and uses the actual energy consumption values ​​corresponding to each group of test data as the prediction target. The actual energy consumption values ​​are the energy consumption values ​​corresponding to the test data collected in advance; the prediction results are evaluated for model accuracy using the mean absolute percentage error (MAPE). When the calculated MAPE is less than the preset MAPE, the energy consumption prediction model training is completed; an energy consumption prediction model is generated that predicts energy consumption values ​​based on the test data; the energy consumption prediction model is a deep belief network model; The training process of the pollution prediction model is the same as that of the energy consumption prediction model, and it is also a deep belief network model; Control and adjust abnormal parameters of equipment according to optimal parameters; Re-collect the process parameters of each device and extract the characteristic parameters, and again determine whether to mark the parameters in the characteristic parameters as abnormal parameters. If abnormal parameters still exist, perform abnormal type analysis and generate early warning instructions.

2. A chemical plant chemical data analysis method based on industrial big data according to claim 1, characterized in that: The temperature includes an inlet temperature and an outlet temperature, the pressure includes an inlet pressure and an outlet pressure, and the flow rate includes an inlet flow rate and an outlet flow rate; The characteristic parameters include temperature characteristic parameters, pressure characteristic parameters and conversion rate; the temperature characteristic parameters include temperature peak value, temperature valley value and temperature difference value; the pressure characteristic parameters include pressure peak value, pressure valley value and pressure difference value; The conversion rate is the product conversion capacity of the equipment; The method for extracting characteristic parameters comprises: The inlet temperature of each device is compared with the outlet temperature, and the temperature with a larger value is taken as the temperature peak value of the corresponding device, and the temperature with a smaller value is taken as the temperature valley value of the corresponding device; the inlet pressure of each device is compared with the outlet pressure, and the pressure with a larger value is taken as the pressure peak value of the corresponding device, and the pressure with a smaller value is taken as the pressure valley value of the corresponding device; the difference between the outlet temperature of each device and the corresponding inlet temperature is taken as the temperature difference of the corresponding device; the difference between the outlet pressure of each device and the corresponding inlet pressure is taken as the pressure difference of the corresponding device; the simulation model corresponding to each device is constructed by using modeling and simulation software, and the simulation model is a mathematical and physical model for numerical description and calculation of the equipment operation process; the inlet flow of each device is input into the corresponding simulation model, and the corresponding outlet flow is predicted and marked as the predicted flow; the predicted flow is divided by the outlet flow collected in real time by the corresponding device to obtain the conversion rate of each device.

3. A chemical plant chemical data analysis method based on industrial big data according to claim 2, characterized in that: The method for determining whether to mark a parameter in the characteristic parameter as an abnormal parameter comprises: Set a different number for each device, and mark the number of each device as the device number; take the characteristic parameters and device number of a device as a set of analysis data, input each set of analysis data into the trained abnormality judgment model, output the corresponding judgment label, and obtain the corresponding judgment result according to the judgment label, and the judgment result includes temperature abnormality, pressure abnormality, normal and abnormal; The training process of the anomaly judgment model includes: Pre-set corresponding judgment results for a group of analysis data, where a is an integer greater than 1, and set different digital labels for different judgment results; mark the digital labels of the judgment results as judgment labels, and convert the analysis data and the corresponding judgment labels into a corresponding set of feature vectors; use each set of feature vectors as input to an abnormality judgment model, and the abnormality judgment model uses a set of predicted judgment labels corresponding to each group of analysis data as output, and uses the actual judgment labels corresponding to each group of analysis data as prediction targets, and the actual judgment labels are pre-collected digital labels of the judgment results corresponding to the analysis data; minimize the sum of the prediction errors of all analysis data as a training target; train the abnormality judgment model until the sum of the prediction errors reaches convergence and the training is stopped; the abnormality judgment model is a deep neural network model; If the judgment result is that the temperature is abnormal, the temperature characteristic parameter is marked as an abnormal parameter; if the judgment result is that the pressure is abnormal, the pressure characteristic parameter is marked as a pressure abnormality; if the judgment result is that all are abnormal, both the temperature characteristic parameter and the pressure characteristic parameter are marked as abnormal; if the judgment result is that all are normal, no characteristic parameter is marked; A corresponding conversion rate threshold is preset for each device, and the conversion rate of each device is compared with the corresponding conversion rate threshold; if the conversion rate is greater than or equal to the conversion rate threshold, the conversion rate is not marked as an abnormal parameter; if the conversion rate is less than the conversion rate threshold, the conversion rate is marked as an abnormal parameter.

4. A chemical plant chemical data analysis method based on industrial big data according to claim 3, characterized in that: In step 6, the method for calculating the clone size includes: F i =min(exp(||X i -X j ||)); In the formula, R i is the clone size of the i-th antibody, Int is the upward rounding function, q is the size of the antibody group, Φ i is the affinity of the ith antibody, min is the minimum function, exp is the natural exponential function, ||X i -X j || is the Euclidean distance between the i-th antibody and the j-th antibody, i≠j, j∈[1,q]; In step 8, the progeny antibody S c The calculation methods include: In the formula, X′ w is the w-th parent antibody, s w is the w-th scaling factor, which is a randomly generated real number and not all w scaling factors are 0.

5. A chemical plant chemical data analysis method based on industrial big data according to claim 4, characterized in that: According to the obtained optimal parameters, the abnormal parameters of the corresponding abnormal equipment are controlled and adjusted respectively; for the equipment whose conversion rate is marked as abnormal parameters, the outlet flow control is adjusted to the corresponding predicted flow.

6. A chemical plant chemical data analysis method based on industrial big data according to claim 5, characterized in that: Methods for performing anomaly type analysis include: If the abnormal parameter is the temperature peak, the abnormal type of the equipment is judged to be a first-level temperature abnormality; If the abnormal parameter is a temperature valley value, the abnormal type of the device is judged to be a secondary temperature abnormality; If the abnormal parameter is the temperature difference, the abnormal type of the equipment is judged to be a level 3 temperature abnormality; If the abnormal parameter is the pressure peak, the abnormal type of the equipment is judged to be a primary pressure abnormality; If the abnormal parameter is a pressure valley value, the abnormal type of the equipment is judged to be a secondary pressure abnormality; If the abnormal parameter is the pressure difference, the abnormal type of the equipment is judged to be a third-level pressure abnormality; Two adjacent devices in the factory production process are regarded as a group of device sets. In each group of device sets, the device connected to the other device at the inlet is marked as a downstream device, and the device connected to the other device at the outlet is marked as an upstream device; the outlet flow of the upstream device in each group of device sets is subtracted from the inlet flow of the downstream device to obtain the flow difference in each group of device sets; a corresponding difference threshold is preset for each group of device sets, and each flow difference is compared with the corresponding difference threshold; if the flow difference is greater than or equal to the difference threshold, the corresponding device set is marked as an abnormal set; if the flow difference is less than the flow threshold, the corresponding device set is not marked as an abnormal set; If the device set is marked as an abnormal set, it is determined that the transmission device between two devices in the device set is abnormal.

7. A chemical plant chemical data analysis method based on industrial big data according to claim 6, characterized in that: Adjust the optimal inlet temperature and the optimal inlet pressure in the optimal parameters corresponding to each device; Connect the simulation models corresponding to each device to obtain the simulation model corresponding to the factory production line, that is, connect and match the input ports and output ports between the simulation models of each device; input the optimal inlet temperature and optimal inlet pressure in the optimal parameters of each device, the outlet temperature and outlet pressure collected in real time, and the flow rate after control adjustment into the corresponding simulation model; adjust the outlet temperature control of each simulation model to the optimal outlet temperature in the optimal parameters of the corresponding device, adjust the outlet pressure control of each simulation model to the optimal outlet pressure in the optimal parameters of the corresponding device, and collect the inlet temperature and inlet pressure of each simulation model; subtract the collected inlet temperature of each simulation model from the corresponding The best inlet temperature in the optimal parameters of the equipment is used to obtain the inlet temperature difference corresponding to each equipment; the inlet temperature of each simulation model collected is subtracted from the best inlet temperature in the optimal parameters of the corresponding equipment to obtain the inlet temperature difference corresponding to each equipment; the best inlet temperature in the optimal parameters corresponding to each equipment is subtracted from the corresponding inlet temperature difference to obtain the inlet regulating temperature; the best inlet pressure in the optimal parameters corresponding to each equipment is subtracted from the corresponding inlet pressure difference to obtain the inlet regulating pressure; the best inlet temperature in the optimal parameters corresponding to each equipment is adjusted to the corresponding inlet regulating temperature, and the best inlet pressure in the optimal parameters corresponding to each equipment is adjusted to the corresponding inlet regulating pressure.

8. A chemical plant chemical data analysis system based on industrial big data, implementing a chemical plant chemical data analysis method based on industrial big data as described in any one of claims 1 to 7, characterized in that: include: Parameter collection module, used to collect the process parameters of each equipment in the factory production process in real time, including temperature, pressure and flow; Parameter processing module, used to process the process parameters of each device and extract characteristic parameters; The abnormality judgment module is used to analyze the characteristic parameters of each device and determine whether to mark the parameters in the characteristic parameters as abnormal parameters; The parameter analysis module is used to analyze the traffic of the device corresponding to the abnormal parameters and obtain the corresponding optimal parameters; Parameter control module, used to control and adjust abnormal parameters of equipment according to optimal parameters; The type analysis module is used to re-collect the process parameters of each device and extract the characteristic parameters, and again determine whether to mark the parameters in the characteristic parameters as abnormal parameters. If abnormal parameters still exist, perform abnormal type analysis and generate early warning instructions.

Images