Energy efficiency-oriented steel plant data flow real-time prediction system and method

Abstract

The invention discloses an energy efficiency-oriented steel mill data flow real-time prediction system and method. The system comprises a data acquisition module, a data processing module, a data storage module, an energy consumption visualization module, an energy efficiency calculation module and an energy efficiency prediction module. The energy consumption data flow of each stage and each device is dynamically displayed based on the predefined steelmaking process, the energy consumption data flow of the steel plant is monitored and simulated in real time, and the advanced machine learning and data analysis technology is integrated for comprehensive statistical analysis, so that real-time calculation and future prediction of the energy efficiency of the steel plant are realized; the problems of incomplete energy efficiency management and low prediction model precision in the prior art are solved. The system can help enterprises to reduce energy consumption and improve production efficiency, and provides a new technical path for intelligent cost management and control of industrial production.

Description

Technical Field

[0001] The present invention belongs to the technical field of integrated energy application and big data computing, and relates to a real-time prediction system and method for data streams of a steel plant oriented to energy efficiency. Background Art

[0002] As raw materials for industrial production, energy and resources occupy an unshakable position in the industrial production process. With the rapid development of industrial technology, the demand for diversified energy and resources is increasing. According to statistics, industrial energy consumption accounts for more than 70% of the total social consumption, resulting in low energy and resource utilization, and the imbalance between energy consumption and ecological protection and economic improvement. With the advancement of Industry 4.0, the manufacturing industry is undergoing an unprecedented transformation. The concepts of intelligent manufacturing and digital factories are constantly being proposed and applied, aiming to improve energy utilization, production efficiency and product quality through the deep integration of information technology and manufacturing technology. How to improve the energy efficiency of the production system and reduce energy consumption and production costs has become an urgent problem to be solved in industrial development.

[0003] Energy efficiency, as an important evaluation indicator of industrial production, is not only related to the operating costs of the factory, but also directly affects environmental protection and sustainable development. Traditional industrial systems often find it difficult to monitor, optimize energy consumption and improve energy efficiency in real time, resulting in low energy utilization. To solve this problem, modern industrial systems make full use of Internet technology and gradually introduce data stream real-time simulation technology. Data stream real-time simulation is a method of simulating and optimizing production systems by real-time collection, processing and analysis of data streams in the production process. This technology can reflect the dynamic changes of industrial production systems in real time. By combining machine learning technologies such as prediction, it can timely identify and solve energy efficiency problems in the production process, thereby achieving efficient use of energy and optimization of production systems.

[0004] The patent "An Industrial Energy Efficiency Management System and Method" with publication number CN109325641A discloses an industrial energy efficiency management system including an acquisition module, a calculation module and a management module. The energy efficiency management of this invention is not comprehensive enough, cannot display the real-time simulation data flow of the industrial process, and cannot predict the energy efficiency of the enterprise in the future. The patent "A Real-time Simulation Method and System for an Integrated Energy System" with publication number CN111832161A will build a simulation model including a heat network and a natural gas network, as well as a power grid model, to form a coupled simulation model. This patent application focuses on combining meteorological data to display the energy consumption data of buildings in real time. It is not suitable for industrial production processes and does not involve future energy efficiency analysis.

[0005] In the research on energy efficiency prediction, the authorized invention patent CN114004423A proposed an energy demand prediction method and system considering regional energy efficiency collaborative optimization. The regional energy consumption data matrix and the regional GDP matrix were input into the deep belief network to train the features and obtain the energy demand prediction value; the authorized patent CN111874182A introduced an energy efficiency prediction control system and method in the hybrid ship scenario. By establishing an energy efficiency prediction speed model, a power prediction model and a power dynamic management model, the recommended speed, required power, engine output power and propeller speed of the ship in the next time period are obtained, which is convenient for the ship to improve energy utilization and sail economically.

[0006] Energy efficiency prediction and simulation technology is of great significance in modern industry, especially in high-energy-consuming industries such as steel mills. By accurately predicting energy efficiency, this technology can help companies optimize production processes, reduce energy consumption, and reduce greenhouse gas emissions, thereby achieving sustainable development goals. In addition, simulation technology can simulate the effects of different production strategies and process improvements in a virtual environment, providing a scientific basis for corporate decision-making. This can not only significantly reduce operating costs, but also improve production efficiency and equipment utilization, and enhance the competitiveness of companies in the market. Therefore, energy efficiency prediction and simulation technology is crucial to promoting industrial energy conservation and emission reduction and improving resource utilization efficiency. A comprehensive analysis of the above research and published patents found that the existing energy efficiency prediction and simulation technology has the following problems:

[0007] 1. Data collection technology with high latency, high cost and high difficulty. Industrial heterogeneous Internet of Things realizes data collection through device intercommunication or unified data format. For example: using gateways to connect heterogeneous networks or protocols requires a large amount of data conversion and transmission, with slow transmission speed and poor real-time performance, and high cost for maintaining gateway devices; using protocol converters to realize conversion of different communication protocols, but this technology cannot adapt to all protocols and cannot collect data on some special protocol devices, and the protocol analysis and conversion calculations of protocol converters are relatively complex, which is prone to low transmission rate and poor performance; using a unified data model to unify various data formats on different devices increases data complexity and management difficulty, which in turn leads to high computing overhead and high collection delay.

[0008] 2. The simulation system is not perfect. Some systems only count the energy consumption data of energy resources in real time, and do not combine the industrial production process to simulate the dynamic data flow process, and cannot provide non-professionals with a clear industrial production process. At the same time, other systems do not involve energy efficiency calculations, especially future energy efficiency prediction research, and cannot provide factories with energy efficiency references to effectively reduce resource consumption and production costs, which is not conducive to ecological protection and sustainable development.

[0009] 3. There are deficiencies in energy efficiency prediction models. In industrial energy efficiency research, energy efficiency prediction models are mainly divided into two categories: traditional dynamic models and data-driven models of intelligent algorithms. However, traditional dynamic models require professional energy dynamics derivation, and modeling is relatively complex and difficult to meet modeling requirements under different conditions, making it impossible to obtain high-precision energy efficiency prediction models; the data-driven model of intelligent algorithms relies on machine learning and deep learning algorithms, which is a black box model with strong nonlinear fitting effects, but this model cannot guarantee high model accuracy and strong generalization ability, and this black box model has poor interpretability.

[0010] Research on real-time simulation and prediction technology for energy efficiency data streams of industrial production systems is still immature. Although existing energy efficiency management systems can collect energy consumption data of production systems through collection equipment, it is still a huge challenge to make full use of the collected energy consumption data to implement complex data analysis to find problems and provide decision-making references. Its data analysis often uses basic descriptive statistics, and data visualization only displays trend analysis such as summary and mean in the form of charts. Moreover, data collection technology faces the disadvantages of high cost and high latency; providing production energy efficiency and future energy efficiency to supervisors in real time is also worth considering. At present, there are some data flow simulation systems involving energy efficiency calculation processes, but they lack the function of predicting future energy efficiency, forming barriers to managing energy resource consumption costs and regulatory cost benefits. Therefore, helping enterprises implement data supervision, reduce production costs, adapt to market demand and achieve green manufacturing are still urgent problems to be solved. Summary of the invention

[0011] In order to solve the above technical problems, the purpose of the present invention is to provide a steel plant data flow real-time prediction system and method oriented to energy efficiency.

[0012] The present invention provides a steel plant data flow real-time prediction system for energy efficiency, comprising: a data acquisition module, a data processing module, a data storage module, an energy consumption visualization module, an energy efficiency calculation module and an energy efficiency prediction module;

[0013] The data acquisition module is used to collect energy consumption data of energy resource materials on each edge equipment in the steelmaking process;

[0014] The data processing module is connected to the data acquisition module and the data storage module respectively, and is used to pre-process the energy consumption data collected from the edge device;

[0015] The data storage module is used to store the pre-processed energy consumption data collected within a period of time into a time series database;

[0016] The energy consumption visualization module is used to visualize the real-time energy consumption data flow of each stage and each device in the predefined steelmaking process, and to perform energy consumption data aggregation and statistical analysis visualization;

[0017] The energy efficiency calculation module is used to obtain energy consumption data within a period of time from the data storage module, and calculate the steelmaking energy efficiency within this period of time according to the energy efficiency calculation formula;

[0018] The energy efficiency prediction module is used to construct a multi-core regularized extreme learning machine prediction model based on improved particle swarm optimization, input energy consumption data and steelmaking energy efficiency over a period of time into the prediction model, and predict the future energy efficiency trend of the steel plant.

[0019] The present invention also provides a steel plant data stream real-time prediction method for energy efficiency, comprising the following steps:

[0020] Step 1: Receive and parse the data collection instructions sent by the server, deploy the data collection agent on the edge device, and create data collection points to complete energy consumption data collection;

[0021] Step 2: The data collection agent reports the energy consumption data, and performs data cleaning, data format normalization, outlier discarding and missing value filling on the original energy consumption data;

[0022] Step 3: Store the preprocessed energy consumption data into the time series database;

[0023] Step 4: Read the energy consumption data in the time series database, visualize the data flow at each stage and each device in the steelmaking process, and summarize and count the energy consumption data;

[0024] Step 5: Read the energy consumption data in the time series database, and solve the steelmaking energy efficiency data of the steel plant within a specific time interval based on the energy efficiency calculation process;

[0025] Step 6: The energy consumption data and steelmaking energy efficiency data are combined into a data set, which is divided into a training set and a test set and transmitted to the energy efficiency prediction module. The multi-core regularized extreme learning machine prediction model based on improved particle swarm optimization is trained to predict the future energy efficiency of the steel plant.

[0026] The present invention provides a real-time prediction system and method for data streams in a steelmaking plant for energy efficiency, which has the following beneficial effects:

[0027] (1) The system functions are relatively comprehensive. The present invention is a data flow simulation system that integrates real-time data collection, data statistical analysis, data flow visualization simulation, energy efficiency calculation and energy efficiency prediction. Compared with the existing industrial data flow simulation system, it has more complete functions;

[0028] (2) The data collection technology is more flexible and simple. The data collection technology in the present invention designs different data collection programs for different protocols, manages these data collection programs uniformly through the agent program, creates data collection programs according to the server instructions, and realizes data collection. Different from the existing collection technology, the present invention does not use the protocol conversion method, does not define a unified data type, and is compatible and adaptable to multiple data protocols;

[0029] (3) Real-time display of industrial production process data. The present invention visualizes the real-time data collected from each stage and each device in the industrial production process, and also includes the summary and statistical analysis of energy consumption data flow over a period of time. Compared with the prior art, this system visualizes the industrial production process data flow in real time, making the industrial production process clearer and more real-time;

[0030] (4) Real-time tracking of existing energy efficiency and future energy efficiency. The energy efficiency calculation module of the present invention uses the collected energy consumption data to solve the energy efficiency of the steel plant over a period of time, and can display the energy efficiency of the production process at any time. At the same time, the energy efficiency prediction model can predict the future energy efficiency trend of the steel plant, which is convenient for relevant personnel of the enterprise to understand the dynamic changes in the energy efficiency of the production process, and adjust the production structure based on the future energy efficiency trend, reduce energy consumption costs, and achieve efficient utilization of energy resources;

[0031] (5) The energy efficiency prediction model is more reasonable. The present invention adopts a multi-core extreme learning machine algorithm based on improved particle swarm optimization to establish an energy efficiency prediction model to predict the future energy efficiency of the factory. The multi-core extreme learning machine has the characteristics of fewer learning parameters, fast training speed, and strong generalization. It is suitable for processing nonlinear problems of large-scale data. At the same time, the particle swarm algorithm is used to solve the network parameters of the extreme learning machine, replacing the random initialization parameters of the extreme learning machine, and three improved schemes are used to optimize the particle swarm algorithm, so that the constructed extreme learning machine network is more reasonable, the prediction model is more generalized, the accuracy is higher, and the prediction effect is better. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 is a schematic diagram of a steel plant data stream real-time prediction system for energy efficiency of the present invention;

[0033] Figure 2 It is a flow chart of a steel plant data stream real-time prediction method for energy efficiency of the present invention;

[0034] Figure 3 It is a schematic diagram of optimization of a multi-core regularized extreme learning machine model based on improved particle swarm optimization of the present invention;

[0035] Figure 4 This is the extreme learning machine network architecture diagram;

[0036] Figure 5It is a flow chart of the multi-core regularized extreme learning machine prediction model based on improved particle swarm optimization of the present invention;

[0037] Figure 6 It is an energy efficiency prediction effect diagram of the multi-core regularized extreme learning machine prediction model based on improved particle swarm optimization of the present invention;

[0038] Figure 7 This is a loss graph comparing the multi-core regularized extreme learning machine prediction model based on improved particle swarm optimization of the present invention with other models. DETAILED DESCRIPTION

[0039] like Figure 1 As shown, the present invention provides a real-time prediction system for data flow of a steel plant for energy efficiency, comprising: a data acquisition module, a data processing module, a data storage module, an energy consumption visualization module, an energy efficiency calculation module and an energy efficiency prediction module.

[0040] The data acquisition module is used to collect energy consumption data of energy resource materials on each edge device in the steelmaking process. The data processing module is connected to the data acquisition module and the data storage module respectively, and is used to pre-process the energy consumption data collected from the edge devices. The data storage module is used to store the pre-processed energy consumption data collected within a period of time into a time series database. The energy consumption visualization module is used to visualize the real-time energy consumption data stream on each stage and each device in the predefined steelmaking process, and to visualize the energy consumption data summary and statistical analysis. The energy efficiency calculation module is used to obtain energy consumption data within a period of time from the data storage module, and calculate the steelmaking energy efficiency within this period of time according to the energy efficiency calculation formula. The energy efficiency prediction module is used to construct a multi-core extreme learning machine prediction model based on improved particle swarm optimization, input the energy consumption data and steelmaking energy efficiency within a period of time into the prediction model, and predict the future energy efficiency trend of the steelmaking plant.

[0041] The data acquisition agent is deployed on the edge device. The server in the data acquisition module sends the data acquisition instructions to the data acquisition agent through the MQTT Broker. The data acquisition agent creates a data channel based on different data channel protocols (i.e., OPCDA and OPC UA), and creates data points under the data channel to complete the data collection tasks on each edge device. The data points represent the sensors set on each edge device of the steelmaking process, which are used to collect energy consumption data at each stage and on each device.

[0042] In the data processing module, the original energy consumption data collected by the data acquisition module is matched with the corresponding data processing logic according to its data type, and data cleaning, data format normalization, outlier discarding and missing value completion are performed to improve the quality of energy consumption data.

[0043] In the energy consumption visualization module, based on the pre-defined steelmaking process, energy consumption data for a period of time or energy consumption data after real-time processing is read from the time series database, the energy consumption data flow at each stage and on each equipment in the steelmaking process is visualized, and statistical charts of energy materials are visualized.

[0044] In the energy efficiency calculation module, energy consumption data over a period of time is obtained from the time series database, mainly including energy resource consumption, electricity consumption, steel production, and recovered heat and gas. Then, the steelmaking energy efficiency data of the steel plant over a period of time is solved based on the energy efficiency calculation formula.

[0045] In the energy efficiency prediction module, the improved particle swarm optimization and multi-core extreme learning machine prediction models are combined to form a multi-core regularized extreme learning machine prediction model based on improved particle swarm optimization. The energy consumption data and steelmaking energy efficiency over a period of time are input into the prediction model to predict the future energy efficiency trend of the steel plant.

[0046] like Figure 2 As shown, a real-time prediction method for steel plant data stream oriented to energy efficiency includes the following steps:

[0047] Step 1: Receive and parse the data collection instructions sent by the server, deploy the data collection agent on the edge device, and create data collection points to complete energy consumption data collection.

[0048] Step 2: The data collection agent reports the energy consumption data, and performs data cleaning, data format normalization, outlier discarding and missing value filling on the original energy consumption data.

[0049] Step 3: Store the preprocessed energy consumption data into the time series database.

[0050] Step 4: Read the energy consumption data in the time series database, visualize the data flow at each stage and on each equipment in the steelmaking process, and summarize and count the energy consumption data.

[0051] Step 5: Read the energy consumption data in the time series database, and solve the steelmaking energy efficiency data of the steel plant within a specific time interval based on the energy efficiency calculation process. The energy efficiency calculation process in step 5 is specifically as follows:

[0052] Step 5.1: Read the energy consumption data in the time series database, including electricity consumption, consumption of various energy resources, steel output, and recovered heat and gas.

[0053] Step 5.2: Calculate the total energy consumption of the steel plant. The total energy consumption of the steelmaking process includes electricity consumption and fuel combustion consumption. The total energy consumption calculation formula is:

[0054] E 总 =P 电力 +F燃料 ×η 燃料转化

[0055] Among them, E 总 is the total energy consumption of the steelmaking process, in kilowatt-hours; P 电力 is the total electricity consumption in the steelmaking process, in kilowatt-hours; F 燃料 It is the consumption of various fuels (such as coal, coke, natural gas, steam and fuel oil) in the steelmaking process, in tons or cubic meters; η 燃料转化 It is the fuel conversion rate, expressed in tons per kilowatt-hour or cubic meters per kilowatt-hour.

[0056] Step 5.3: Calculate the net energy consumption of the steelmaking plant. The net energy consumption of the steelmaking process refers to the energy resource consumption excluding the recovered energy or recovered gas. The calculation formula for the net energy consumption is:

[0057] E 净 =E 总 -E 回收热能 -E 回收气体

[0058] Among them, E 净 is the net energy consumption of the steelmaking process, in kilowatt-hours; E 回收热能 The unit of waste heat for recycling is kWh; E 回收气体 The energy generated by the recycled exhaust gas, in kilowatt-hours.

[0059] Step 5.4: Energy efficiency refers to the ratio of the actual energy resource consumption in the steelmaking process, also known as the inverse of energy consumption per unit output. The formula for calculating steelmaking energy efficiency is:

[0060]

[0061] in, is energy efficiency, in tons per kilowatt-hour, T 产量 It is the total steel production recorded in a period of time, in tons.

[0062] Step 6: The energy consumption data and steelmaking energy efficiency data are combined into a data set, which is divided into a training set and a test set and transmitted to the energy efficiency prediction module. The multi-core regularized extreme learning machine prediction model IPSO-MKELM based on improved particle swarm optimization is trained to predict the future energy efficiency of the steel plant.

[0063] like Figure 3 The figure shows the optimization diagram of the multi-core regularized extreme learning machine prediction model based on improved particle swarm optimization. In order to facilitate the understanding of the algorithm used in the prediction model, the principles of extreme learning machine and multi-core extreme learning machine are briefly introduced before building the prediction model.

[0064] Extreme Learning Machine (ELM) is a single hidden layer feedforward neural network. Unlike traditional neural networks, it does not need to repeatedly iterate to calculate weights, but directly obtains weights by solving the Moore-Penrose generalized inverse matrix.

[0065] An ELM network with L hidden layer nodes is as follows: Figure 4 , assuming there are N samples (x i ,y i ), i=1,2,3...,N,x i is the data set, y i To correspond to the sample label, the hidden layer output of the extreme learning machine is as follows:

[0066] h j (x i ) = g j (w j ,b j ,x i ) = g j (w j x i +b j )

[0067] Among them, h j (x i ) is the output of the hidden layer, g(·) is the activation function, w j and b j weights and biases respectively.

[0068] The network output layer expression of the extreme learning machine is as follows:

[0069]

[0070] Among them, i is the network output, L is the number of hidden layer nodes, β j is the weight matrix of the hidden layer neurons and the output layer.

[0071] The goal of the extreme learning machine is to ensure that the training error is minimized. By randomly initializing weights and biases, the objective function is solved to obtain the weight matrix. The objective function and weight matrix are as follows:

[0072] min||Hβ-Y|| 2

[0073]

[0074] Where H = [h1(x i ),h2(x i ),...,h L (xi )] is the hidden layer output matrix, satisfying H∈R N×L , β=[β1,β2,...,β L ] T is the weight matrix, Y=[y1,...y N ] is the sample label matrix, H -1 is the Moore-Penrose generalized inverse matrix of H.

[0075] By introducing the regularization coefficient, we get the regularized extreme learning machine (RELM), which aims to ensure that the training error is minimized while the output weight is minimized. The objective function and output weight matrix of RELM are as follows:

[0076]

[0077] Where C is the regularization coefficient. When N ≥ L, I∈R N×N For the unit array.

[0078] Finally, the hidden layer output is multiplied by the output weight matrix to obtain the network output of RELM, as shown below:

[0079]

[0080] The linear parameter solution process of ELM is difficult to solve the problem that the original input data is nonlinear. Introducing the kernel function into the extreme learning machine can map complex nonlinear data into a high-dimensional space, making it linearly separable and significantly reducing the amount of computation for processing large-scale data sets.

[0081] Based on the above, if Figure 5 As shown, the specific process of constructing a multi-core regularized extreme learning machine prediction model based on improved particle swarm optimization and training the model is as follows:

[0082] Step 6.1: Introduce the mixed Gaussian kernel function and the second-order polynomial kernel function into the regularized extreme learning machine to obtain a multi-kernel regularized extreme learning machine model;

[0083] The mixed Gaussian kernel function and the second-order polynomial kernel function are as follows:

[0084]

[0085] Among them, C1 and C2 are the weights of Gaussian kernel function and second-order polynomial kernel function respectively; satisfying C1+C2=1, is the Gaussian kernel function, σ is the kernel function bandwidth, (ax i x j +c) 2is a second-order polynomial kernel function, a is the scaling factor of the inner product of the control vector, and c is the bias term that controls the translation of the model in the feature space; N samples (x i ,y i ), i=1,2,3...,N,x i is the energy consumption data, y i is the steelmaking energy efficiency data, that is, x i The corresponding sample labels;

[0086] Using the hybrid kernel function as the regularized extreme learning machine kernel function, the kernel matrix is ​​obtained as follows:

[0087]

[0088] Where H = [h1(x i ),h2(x i ),...,h L (x i )] is the hidden layer output matrix, h1(x i ) is the output of the first neuron in the hidden layer, and L is the number of nodes in the hidden layer.

[0089] The kernel matrix is ​​introduced into the regularized extreme learning machine to obtain a multi-core regularized extreme learning machine model. The output weight matrix of the multi-core regularized extreme learning machine model is as follows:

[0090]

[0091] Among them, C is the regularization coefficient, Y=[y1,...y N ], when N ≥ L, I∈R N×N For the unit array.

[0092] The hidden layer output is multiplied by the output weight matrix to obtain the output of the multi-core regularized extreme learning machine model, as shown below:

[0093]

[0094] Step 6.2: Determine the particles and initialize the parameters. Set the hidden layer weights w and bias b of the multi-core regularized extreme learning machine model as particles, and the parameters σ, C1, a, and c. Initialize the particle population size M and the maximum number of iterations k. max And the search space dimension D, the initial inertia weight ω is obtained by the inertia weight optimization scheme, and its optimization formula is as follows:

[0095]

[0096] Where k is the number of iterations, N(0,1) generates random numbers that follow a normal distribution, and ω min is the minimum inertia weight, ω maxis the maximum inertia weight.

[0097] The individual learning factor c1 and social learning factor c2 are initialized by the learning factor optimization scheme, and the optimization formula is as follows:

[0098] c1=0.7-(1.61×cos(πω))

[0099] c2=2.2-c1

[0100] Step 6.3: Initialize the particle speed and position, randomly generate a D-dimensional vector in [0,1] as the particle position, the speed of the ith particle and the position of the ith particle are as follows:

[0101] V i =[V i,1 ,V i,2 ,...,V i,D ] T

[0102] X i =[X i，1 ,X i，2 ,...,X i,D ] T

[0103] Then, the chaotic mapping scheme is used to optimize the positions X of all initialized particles. i .

[0104] In specific implementation, the chaotic mapping scheme in step 6.3 includes the following steps:

[0105] Step 6.3.1: Randomly generate a D-dimensional vector whose elements are in the range [0,1].

[0106] Step 6.3.2: Use Logistic chaos mapping to map the D-dimensional vector to the initial position of the population particles. The formula is as follows:

[0107] X s+1 =μX s (1-X s )

[0108] Among them, μ is the chaotic mapping parameter, s is the number of chaotic mapping iterations, X s is the particle position at the sth iteration.

[0109] Step 6.4: The energy consumption data and steelmaking energy efficiency data are combined into a data set and divided into a training set and a test set.

[0110] Step 6.5: Input the training set into the multi-core regularized extreme learning machine model MKRELM to train the MKRELM model.

[0111] Step 6.6: Calculate the fitness and determine the individual extreme value and the global optimal solution. The initial position of the particle is used as the individual extreme value p best,i The root mean square error is used as the fitness function, and the fitness values ​​of all particles are calculated based on the multi-core regularized extreme learning machine model MKRELM, and the particle position with the highest fitness value among all particles is selected as the initial global optimal solution g best , the expressions of individual extreme value and global optimal solution are as follows:

[0112] p best,i =[p i1 ,p i2 ,...,p iD ] T

[0113] g best =[g i1 ,g i2 ,...,g iD ] T

[0114] Step 6.7: Iterate to find the global optimal solution. Perform iterative optimization operation on all particles to obtain the global optimal solution.

[0115] The iterative optimization operation in step 6.7 includes the following steps:

[0116] Step 6.7.1: Update the particle velocity and position, and ensure that the particle position is still within the established search space. The particle velocity and position update formulas are as follows:

[0117]

[0118] Among them, r1 and r2 are random numbers in [0,1]. is the individual extreme value of the kth iteration, is the global optimal solution of the kth iteration, is the position of the i-th particle at the k-th iteration, V i k is the velocity of the ith particle at the kth iteration, is the vector from the current position of the particle to its optimal position, It is the vector pointing from the current position of the particle to the optimal position of the group.

[0119] Step 6.7.2: Calculate the fitness value of the particle. Based on the fitness function, namely the root mean square error (RMSE), calculate the fitness value of all particles after the update. The calculation formula is as follows:

[0120]

[0121] in, is the predicted value, y i is the dataset label.

[0122] Step 6.7.3: Update the individual extreme value and the global optimal solution. If the fitness value is better than the individual extreme value, update the individual extreme value p best,i ; If the fitness value is better than the fitness value of the global optimal solution, then update the global optimal solution g best .

[0123] Step 6.7.4: Determine the termination condition. If the maximum number of iterations is reached or the fitness threshold is reached, terminate the iteration; otherwise, go to step 6.7.1 to continue the iteration.

[0124] Step 6.8: Obtain the optimal particle value. Obtain the optimal value of all particles by iteration, that is, obtain the optimal values ​​of the hidden layer weights and biases of the MKRELM model and the parameters σ, C1, a, and c;

[0125] Step 6.9: Input the test set into the MKRELM model, test the model, and obtain the prediction results.

[0126] The daily energy efficiency data of the steel plant from January 1, 2024 to November 3, 2024 was simulated by Python, and divided into training set and test set in a ratio of 8:2 in chronological order. The extreme learning machine was trained with the training set to determine the model parameters to be estimated, and the model was inferred with the test set. Finally, the energy efficiency of the steel plant in the next two weeks was predicted. The energy efficiency prediction effect is as follows: Figure 6 As shown. The figure shows the test set fitting effect from October 3 to November 3. The data of 88 days before October 3 is used as input data and input into the corresponding prediction model to predict the energy efficiency value from October 3 to November 3. It is observed that the IPSO-MKELM model result (dashed five-pointed star) proposed in this paper is closest to the test set (black solid line), while the prediction fitting effect of the comparison model PSO-MKELM (square-dashed line), IPSO-rbf (triangle-dashed line), and IPSO-poly (circle-dashed line) is poor. The figure shows the predicted energy efficiency of the next two weeks from November 4 to November 18. The data of 88 days before November 4 is input into the model to predict the future energy efficiency. Figure 7 The root mean square error of the present invention and the three models is intuitively compared, and it is found that the root mean square error of the proposed IPSO-MKELM is the smallest, that is, MSE=0.00568, which shows that the proposed prediction model has high accuracy.

[0127] Algorithm description in the figure: IPSO-MKELM (five-pointed star-dashed line) is a multi-core regularized extreme learning machine prediction algorithm based on improved particle swarm optimization proposed in the present invention; PSO-MKELM (square-dashed line) is an unimproved particle swarm algorithm, which is a multi-core regularized extreme learning machine prediction algorithm based on the original particle swarm algorithm optimization; IPSO-rbf (triangle-dashed line) represents a prediction algorithm that uses only Gaussian kernel function as kernel function and does not use mixed kernel function; IPSO-poly (circle-dashed line) represents a prediction algorithm that uses only linear kernel function as kernel function and does not use mixed kernel function.

[0128] The above description is only a preferred embodiment of the present invention and is not intended to limit the concept of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A real-time prediction system for steel plant data streams for energy efficiency, characterized in that: include: Data acquisition module, data processing module, data storage module, energy consumption visualization module, energy efficiency calculation module and energy efficiency prediction module; The data acquisition module is used to collect energy consumption data of energy resource materials on each edge equipment in the steelmaking process; The data processing module is connected to the data acquisition module and the data storage module respectively, and is used to pre-process the energy consumption data collected from the edge device; The data storage module is used to store the pre-processed energy consumption data collected within a period of time into a time series database; The energy consumption visualization module is used to visualize the real-time energy consumption data flow of each stage and each device in the predefined steelmaking process, and to perform energy consumption data aggregation and statistical analysis visualization; The energy efficiency calculation module is used to obtain energy consumption data within a period of time from the data storage module, and calculate the steelmaking energy efficiency within this period of time according to the energy efficiency calculation formula; The energy efficiency prediction module is used to construct a multi-core regularized extreme learning machine prediction model based on improved particle swarm optimization, input energy consumption data and steelmaking energy efficiency over a period of time into the prediction model, and predict the future energy efficiency trend of the steel plant.

2. The energy efficiency-oriented steel plant data stream real-time prediction system according to claim 1, characterized in that: The data collection agent is deployed on the edge device. The server in the data collection module sends the data collection instruction to the data collection agent through MQTTBroker. The data collection agent creates a data channel based on different data channel protocols, creates data points under the data channel, and completes the data collection task on each edge device. The data points represent the sensors installed on the equipment at each end of the steelmaking process, which are used to collect energy consumption data at each stage and on each device.

3. The energy efficiency-oriented steelmaking plant data stream real-time prediction system according to claim 1, characterized in that: In the data processing module, the original energy consumption data collected by the data acquisition module is matched with the corresponding data processing logic according to its data type, and data cleaning, data format normalization, outlier discarding and missing value completion are performed to improve the quality of energy consumption data.

4. The energy efficiency-oriented steelmaking plant data stream real-time prediction system according to claim 1, characterized in that: In the energy consumption visualization module, based on the pre-defined steelmaking process, energy consumption data for a period of time or energy consumption data after real-time processing is read from the time series database, the energy consumption data flow at each stage and on each equipment in the steelmaking process is visualized, and statistical charts of energy materials are visualized.

5. The real-time prediction system for energy efficiency-oriented steelmaking plant data stream according to claim 1, characterized in that In the energy efficiency calculation module, energy consumption data over a period of time is obtained from the time series database, mainly including energy resource consumption, electricity consumption, steel production, and recovered heat and gas. Then, the steelmaking energy efficiency data of the steel plant over a period of time is solved based on the energy efficiency calculation formula.

6. A real-time prediction method for steel plant data streams oriented to energy efficiency, characterized in that: The following steps are involved: Step 1: Receive and parse the data collection instructions sent by the server, deploy the data collection agent on the edge device, and create data collection points to complete energy consumption data collection; Step 2: The data collection agent reports the energy consumption data, and performs data cleaning, data format normalization, outlier discarding and missing value filling on the original energy consumption data; Step 3: Store the preprocessed energy consumption data into the time series database; Step 4: Read the energy consumption data in the time series database, visualize the data flow at each stage and each device in the steelmaking process, and summarize and count the energy consumption data; Step 5: Read the energy consumption data in the time series database, and solve the steelmaking energy efficiency data of the steel plant within a specific time interval based on the energy efficiency calculation process; Step 6: The energy consumption data and steelmaking energy efficiency data are combined into a data set, which is divided into a training set and a test set and transmitted to the energy efficiency prediction module. The multi-core regularized extreme learning machine prediction model based on improved particle swarm optimization is trained to predict the future energy efficiency of the steel plant.

7. The method for real-time prediction of data stream in a steel plant for energy efficiency according to claim 6, characterized in that: The energy efficiency calculation process in step 5 is specifically as follows: Step 5.1: Read the energy consumption data in the time series database, including electricity consumption, consumption of various energy resources, steel output, and recovered heat and gas; Step 5.2: Calculate the total energy consumption of the steel plant. The total energy consumption of the steelmaking process includes electricity consumption and fuel combustion consumption. The total energy consumption calculation formula is: E 总 =P 电力 +F 燃料 ×η 燃料转化 Among them, E 总 is the total energy consumption of the steelmaking process, in kWh; P 电力 is the total electricity consumption in the steelmaking process, in kilowatt-hours; F 燃料 It is various fuels in the steelmaking process, measured in tons or cubic meters; η 燃料转化 is the fuel conversion rate, expressed in tons per kilowatt-hour or cubic meters per kilowatt-hour; Step 5.3: Calculate the net energy consumption of the steelmaking plant. The net energy consumption of the steelmaking process refers to the energy resource consumption excluding the recovered energy or recovered gas. The calculation formula for the net energy consumption is: AND 净 =And 总 -AND 回收热能 -AND 回收气体 Among them, E 净 is the net energy consumption of the steelmaking process, in kilowatt-hours; E 回收热能 The unit of waste heat for recycling is kWh; E 回收气体 The energy generated by the recycled exhaust gas, in kilowatt-hours; Step 5.4: Energy efficiency refers to the ratio of the actual energy resource consumption in the steelmaking process, also known as the inverse of energy consumption per unit output. The formula for calculating steelmaking energy efficiency is: in, is energy efficiency, in tons per kilowatt-hour, T 产量 It is the total steel production recorded in a period of time, in tons.

8. The method for real-time prediction of data stream in a steel plant for energy efficiency according to claim 6, characterized in that: The specific process of step 6 is as follows: Step 6.1: Introduce the mixed Gaussian kernel function and the second-order polynomial kernel function into the regularized extreme learning machine to obtain a multi-kernel regularized extreme learning machine model; The mixed Gaussian kernel function and the second-order polynomial kernel function are as follows: Among them, C1 and C2 are the weights of Gaussian kernel function and second-order polynomial kernel function respectively; satisfying C1+C2=1, is the Gaussian kernel function, σ is the kernel function bandwidth, (ax i x j +c) 2 is a second-order polynomial kernel function, a is the scaling factor of the inner product of the control vector, and c is the bias term that controls the translation of the model in the feature space; N samples (x i ,y i ), i=1,2,3...,N,x i is the energy consumption data, y i is the steelmaking energy efficiency data, that is, x i The corresponding sample labels; Using the hybrid kernel function as the regularized extreme learning machine kernel function, the kernel matrix is ​​obtained as follows: Where H = [h1(x i ),h2(x i ),...,h L (x i )] is the hidden layer output matrix, h1(x i ) is the output of the first neuron in the hidden layer, and L is the number of nodes in the hidden layer; The kernel matrix is ​​introduced into the regularized extreme learning machine to obtain a multi-core regularized extreme learning machine model. The output weight matrix of the multi-core regularized extreme learning machine model is as follows: Among them, C is the regularization coefficient, Y=[y1,...,y N ], when N ≥ L, I∈R N×N is the unit matrix; The hidden layer output is multiplied by the output weight matrix to obtain the output of the multi-core regularized extreme learning machine model, as shown below: Step 6.2: Determine the particles and initialize the parameters. Set the hidden layer weights w and bias b of the multi-core regularized extreme learning machine model as particles, and the parameters σ, C1, a, and c. Initialize the particle population size M and the maximum number of iterations k. max And the search space dimension D, the initial inertia weight ω is obtained by the inertia weight optimization scheme, and its optimization formula is as follows: Where k is the number of iterations, N(0,1) generates random numbers that follow a normal distribution, and ω min is the minimum inertia weight, ω max is the maximum inertia weight; The individual learning factor c1 and social learning factor c2 are initialized by the learning factor optimization scheme, and the optimization formula is as follows: c1=0.7-(1.61×cos(πω)) c2=2.2-c1 Step 6.3: Initialize the particle speed and position, randomly generate a D-dimensional vector in [0,1] as the particle position, the speed of the ith particle and the position of the ith particle are as follows: V i =[V i,1 ,V i,2 ,...,V i,D ] T X i =[X i，1 ,X i，2 ,...,X i,D ] T Then, the chaotic mapping scheme is used to optimize the positions X of all initialized particles. i ; Step 6.4: The energy consumption data and steelmaking energy efficiency data are used to form a data set, and divided into a training set and a test set; Step 6.5: Input the training set into the multi-core regularized extreme learning machine model to train the multi-core regularized extreme learning machine model; Step 6.6: Calculate the fitness and determine the individual extreme value and the global optimal solution. The initial position of the particle is used as the individual extreme value p best,i , taking the root mean square error as the fitness function, the fitness values ​​of all particles are calculated based on the multi-core regularized extreme learning machine model, and the particle position with the highest fitness value among all particles is selected as the initial global optimal solution g best , the expressions of individual extreme value and global optimal solution are as follows: p best,i =[p i1 ,p i2 ,...,p iD ] T g best =[g i1 ,g i2 ,...,g iD ] T Step 6.7: Iterate to find the global optimal solution, perform iterative optimization operation on all particles to obtain the global optimal solution; Step 6.8: Obtain the optimal particle value. Obtain the optimal value of all particles by iteration, that is, obtain the optimal values ​​of the hidden layer weights and biases of the multi-core regularized extreme learning machine model and the parameters σ, C1, a, and c; Step 6.9: Input the test set into the multi-core regularized extreme learning machine model, test the model, and obtain the prediction results.

9. The method for real-time prediction of data streams in a steel plant for energy efficiency according to claim 8, characterized in that: The chaotic mapping scheme in step 6.3 comprises the following steps: Step 6.3.1: Randomly generate a D-dimensional vector whose elements are in the range [0,1]; Step 6.3.2: Use Logistic chaos mapping to map the D-dimensional vector to the initial position of the population particles. The formula is as follows: X s+1 =μX s (1-X s ) Among them, μ is the chaotic mapping parameter, s is the number of chaotic mapping iterations, X s is the particle position at the sth iteration.

10. The method for real-time prediction of data stream in a steel plant for energy efficiency according to claim 8, characterized in that: The iterative optimization operation in step 6.7 includes the following steps: Step 6.7.1: Update the particle velocity and position, and ensure that the particle position is still within the established search space. The particle velocity and position update formulas are as follows: Among them, r1 and r2 are random numbers in [0,1]. is the individual extreme value of the kth iteration, is the global optimal solution of the kth iteration, is the position of the i-th particle at the k-th iteration, is the velocity of the ith particle at the kth iteration, is the vector from the current position of the particle to its optimal position, It is the vector pointing from the current position of the particle to the optimal position of the group; Step 6.7.2: Calculate the fitness value of the particles. Based on the fitness function, i.e., the root mean square error, calculate the fitness value of all particles after the update. The calculation formula is as follows: in, is the predicted value, y i is the dataset label; Step 6.7.3: Update the individual extreme value and the global optimal solution. If the fitness value is better than the individual extreme value, update the individual extreme value p best,i ; If the fitness value is better than the fitness value of the global optimal solution, then update the global optimal solution g best ; Step 6.7.4: Determine the termination condition. If the maximum number of iterations is reached or the fitness threshold is reached, terminate the iteration; otherwise, go to step 6.7.1 to continue the iteration.

Images