Smart energy management method and system based on industrial internet

Abstract

The invention relates to the technical field of industrial internet, in particular to a smart energy management method and system based on the industrial internet, and the method comprises the following steps: S1, data collection: collecting energy consumption data and environment data in a production process in real time; s2, data fusion: carrying out fusion processing on multi-source data by adopting a data fusion technology; s3, distributed energy optimization: optimizing the energy consumption of each node on the distributed nodes in real time by using an edge computing technology; s4, intelligent load balance optimization: dynamically adjusting the operation load of each industrial production device; s5, multi-level fault prediction: performing fault prediction on the industrial production equipment; and S6, control strategy generation: generating a self-adaptive control strategy in combination with the fault prediction result of the industrial production equipment. According to the method, the lowest energy consumption of the optimization result is ensured, the load is uniformly distributed, local overload and energy waste are avoided, and the overall energy utilization efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of industrial Internet technology, and in particular to a smart energy management method and system based on the industrial Internet. Background Art

[0002] With the rapid development of industrial Internet and intelligent manufacturing, the intelligence and optimization of energy management in industrial production have become an important way for enterprises to enhance their competitiveness. Traditional energy management methods usually rely on manual monitoring and regular maintenance, which are not only inefficient, but also prone to problems such as incomplete data and untimely response. It is difficult to meet the modern industrial production's demand for efficient and accurate energy management. By introducing Internet of Things technology, edge computing and big data analysis technology, real-time monitoring, analysis and optimization of energy consumption and environmental data can be achieved, thereby improving energy utilization efficiency and reducing production costs.

[0003] However, existing technologies still have many shortcomings in intelligent energy management. First, data collection and fusion technology is not perfect, resulting in incomplete and inaccurate collected data, affecting subsequent analysis and optimization. Second, distributed energy optimization and load balancing adjustment lack real-time and dynamic characteristics, and cannot respond to the ever-changing energy consumption demand and equipment status in the industrial production process in a timely manner. In addition, fault prediction technology relies on a single data source and simple analysis methods, and cannot accurately warn of potential faults, resulting in long equipment downtime, high maintenance costs, and low production efficiency.

[0004] In view of the shortcomings of the existing technology, the present invention proposes a smart energy management method based on the industrial Internet, which can realize real-time monitoring, dynamic adjustment and accurate prediction of industrial production equipment, optimize energy distribution, prevent potential failures, ensure stable operation of the system, and significantly improve the overall energy utilization efficiency and economic benefits. Summary of the invention

[0005] Based on the above objectives, the present invention provides a smart energy management method and system based on the Industrial Internet.

[0006] The smart energy management method based on the industrial Internet includes the following steps:

[0007] S1, data collection: through the multimodal sensors deployed in industrial production equipment and environment, real-time collection of energy consumption data and environmental data in the production process, including electricity, heat, gas fuel, temperature, humidity, and air pressure;

[0008] S2, data fusion: the collected energy consumption data and environmental data are fused by using data fusion technology to perform multi-source data fusion processing;

[0009] S3, distributed energy optimization: Based on the results of data fusion, edge computing technology is used to optimize the energy consumption of each node in real time on distributed nodes;

[0010] S4, intelligent load balancing optimization: Based on the optimized energy consumption, dynamically adjust the operating load of each industrial production equipment to achieve optimal distribution of energy load, including:

[0011] S41, data monitoring: real-time monitoring of energy consumption data of each industrial production equipment after optimization;

[0012] S42, load forecasting: Based on the energy consumption data of historical monitoring and real-time monitoring, an intelligent load forecasting model is used to predict the load demand of each industrial production equipment;

[0013] S43, dynamic adjustment: according to the load demand prediction results, the load balancing optimization algorithm is used to dynamically adjust the operating load of each industrial production equipment;

[0014] S5, multi-level fault prediction: Combined with the results of the optimal allocation of energy loads, fault prediction is performed on industrial production equipment, potential faults are warned in advance, and corresponding maintenance plans are formulated;

[0015] S6, control strategy generation: Combined with the fault prediction results of industrial production equipment, generate adaptive control strategies, including equipment operation scheduling, reasonable energy allocation, and production process optimization and adjustment.

[0016] Optionally, the data collection in S1 includes:

[0017] S11, power data collection: Deploy power sensors on industrial production equipment to collect real-time power data consumed during equipment operation;

[0018] S12, thermal energy data collection: Deploy thermal energy sensors in the heating units of industrial production equipment to monitor and collect thermal energy data generated and consumed during equipment operation in real time;

[0019] S13, gas fuel data collection: deploy gas fuel sensors in industrial production environments to collect real-time data on gas fuel consumed by gas boilers or gas equipment;

[0020] S14, environmental temperature and humidity data collection: Deploy temperature sensors and humidity sensors in industrial production environments to monitor and collect environmental temperature and humidity data in real time;

[0021] S15, air pressure data collection: deploy air pressure sensors in industrial production environments to monitor and collect ambient air pressure data in real time.

[0022] Optionally, the data fusion in S2 includes:

[0023] S21, data preprocessing: preliminary processing of the collected energy consumption data and environmental data, including data cleaning, denoising and normalization;

[0024] S22, fusion data: use weighted average fusion algorithm to fuse multi-source data;

[0025] S23, data association analysis: Based on the actual situation of the production process, the fused data is analyzed by calculating the Pearson correlation coefficient to extract key features and patterns.

[0026] Optionally, the distributed energy optimization in S3 includes:

[0027] S31, Edge Node Deployment: Deploy edge computing devices on various distributed nodes of industrial production to process and optimize local energy consumption data in real time;

[0028] S32, data input: input the results of data fusion to each edge node, including consumption data of electricity, heat energy, gas fuel and environmental data;

[0029] S33, optimization target setting: set the optimization target function to minimize energy consumption while meeting production requirements and equipment operation constraints. The optimization target function F is expressed as:

[0030]

[0031] Among them, E i represents the energy consumption of the i-th node, n is the number of nodes;

[0032] S34, Constraints: Considering the production requirements of each node and the equipment operation restrictions, the constraint is set as the maximum power P of the equipment. max and minimum power P min , expressed as:

[0033] P min,i ≤P i ≤P max,i ;

[0034] Among them, P i is the actual power of the i-th node;

[0035] S35, real-time optimization calculation: Use edge computing devices to perform real-time optimization calculations, use linear programming algorithms to solve the optimization objective function, and dynamically adjust the energy consumption distribution of each node under the premise of meeting the constraints, expressed as:

[0036]

[0037] Among them, a iand b i is the optimization coefficient;

[0038] S36, result application: applying the result of the optimization calculation to each node to achieve real-time optimization of the energy consumption of each node.

[0039] Optionally, the data monitoring in S41 includes:

[0040] S411, Sensor deployment: Deploy various types of sensors on various industrial production equipment, including power sensors, thermal energy sensors, and gas fuel sensors, to monitor the energy consumption of the equipment in real time after optimization;

[0041] S412, data collection frequency setting: setting the collection frequency of the electric energy, thermal energy and gas fuel sensors according to the operating characteristics and production requirements of the industrial production equipment.

[0042] Optionally, the intelligent load forecasting model in S42 adopts a Gaussian process regression (GPR) model, and the Gaussian process regression (GPR) model includes:

[0043] S421, define input variables and output variables: Input variable X t Contains historical energy consumption data and environmental data, output variable y t is the actual energy consumption data, expressed as:

[0044] X t =[E t-1 ,E t-2 ,...,E t-n ,θ t ];

[0045] Among them, E t-1 、E t-2 、E t-n Respectively represent the historical energy consumption data at time t-1, t-2, and tn, θ t represents the environmental data at time t;

[0046] S422, define an adaptive kernel function: Based on the characteristics of different input variables, design an adaptive kernel function k, expressed as:

[0047]

[0048] Among them, X i and X j is the input variable, is the signal variance, l d is the length scale of the dth dimension, is the noise variance, δ ij is the Kronecker delta function;

[0049] S423, constructing a covariance matrix: constructing a covariance matrix K according to the adaptive kernel function, expressed as:

[0050]

[0051] Where K(X,X′) is the covariance matrix;

[0052] S424, predictive distribution: predict new energy consumption data y based on the principle of Gaussian process regression * The distribution of is expressed as:

[0053]

[0054] cov(f * )=K(X * ,X * )-K(X * ,X)K(X,X) -1 K(X,X * );

[0055] in, is the predicted mean, cov(f * ) is the prediction covariance;

[0056] S425, Dynamically update model: Dynamically update the Gaussian process regression model according to real-time data, expressed as:

[0057]

[0058] Among them, θ represents the model parameters, α t is the adaptive learning rate at time t, is the gradient of the loss function L with respect to the parameter θ.

[0059] Optionally, the load balancing optimization algorithm in S43 adopts a particle swarm optimization algorithm (PSO), and the particle swarm optimization algorithm (PSO) includes:

[0060] S431, Initialization: Initialize the particle swarm, including the position and velocity of the particles, set the swarm size N, the initial inertia weight w, and the learning factor c 1 and c 2 , maximum number of iterations T, current number of iterations t, expressed as:

[0061] Particle position initialization: x i (0)~U(x min ,x max );

[0062] Particle velocity initialization: v i (0)~U(v min ,vmax );

[0063] Among them, U represents uniform distribution, x min and x max are the minimum and maximum values ​​of the position, respectively, v min and v max are the minimum and maximum values ​​of the speed respectively;

[0064] S432, velocity update: based on the particle's own historical best position p i (t) and the global optimal position g(t), introduce the adaptive inertia weight w(t), and update the particle velocity, which is expressed as:

[0065]

[0066] v i (t+1)=w(t)·v i (t)+c 1 ·r 1 ·(p i (t)-x i (t))+c 2 ·r 2 ·(g(t)-x i (t));

[0067] Among them, w max is the initial inertia weight, w min is the final inertia weight, T is the maximum number of iterations, t is the current number of iterations, r 1 and r 2 is a random number, c 1 and c 2 is the learning factor, x i (t) is the position of particle i at time t, v i (t+1) is the velocity of particle i at time t+1, v i (t) is the velocity of particle i at time t;

[0068] S433, position update: Update the particle's position according to the new velocity, expressed as:

[0069] x i (t+1)=x i (t)+v i (t+1);

[0070] Among them, x i (t+1) is the position of particle i at time t+1;

[0071] S434, fitness evaluation: calculate the fitness value f(x i(t+1)), a multi-objective fitness function is used to comprehensively consider energy consumption and load balance, which is expressed as:

[0072] f(x i )=α·E i +β·B i ;

[0073]

[0074] Among them, E i is the energy consumption of the ith device, B i is the load balancing metric, P j is the actual power of the jth device, α and β are weight coefficients;

[0075] Update the best historical position of the particle:

[0076] Update the global best position: g(t+1)=min i f(p i (t+1));

[0077] S435, Iteration: Repeat the speed update to fitness evaluation until the stopping condition is reached (such as the maximum number of iterations T or the fitness function converges).

[0078] Optionally, the multi-level fault prediction in S5 includes:

[0079] S51, data collection and preprocessing: Collect energy consumption data and environmental data from industrial production equipment, and perform data cleaning, denoising and normalization;

[0080] S52, feature extraction: A convolutional neural network (CNN) model is used to extract fault-related features from the preprocessed energy consumption data and environmental data, expressed as:

[0081] Convolutional Layer:

[0082] Where X is the input data, W is the convolution kernel weight, b is the bias, σ is the activation function, F is the convolution layer output feature map, i, j are the spatial position indexes, k is the feature map index, m, n, p, M, N, P are the convolution kernel sizes;

[0083] Pooling layer:

[0084] Among them, Q is the output feature map of the pooling layer, a, b, A, B are the pooling window sizes;

[0085] S53, fault classification and prediction: Use convolutional neural network (CNN) model for fault classification and prediction, expressed as:

[0086] Flatten a vector: h = flatten(Q);

[0087] First fully connected layer: y 1 =σ(W 1 ·h+b 1 );

[0088] Second fully connected layer (output layer):

[0089] Among them, W 1 is the weight matrix, b 1 is the bias vector, y 1 is the output vector, W 2 is the weight matrix, b 2 is the bias vector, is the output probability vector;

[0090] S54, fault probability calculation: The output results of the convolutional neural network (CNN) model are used to calculate the probability of each fault occurring, expressed as:

[0091]

[0092] where z is the eigenvector, is the probability vector of failure occurrence;

[0093] S55, maintenance plan formulation: Based on the fault prediction results and the probability of fault occurrence, potential faults are warned in advance and corresponding maintenance plans are formulated.

[0094] Optionally, the control strategy generation in S6 includes:

[0095] S61, equipment operation scheduling: according to the equipment fault prediction results, adjust the equipment operation time and load, expressed as:

[0096] S i (t+1)=S i (t)+Δt·(1-P i (fault));

[0097] Among them, S i (t) is the scheduling state of device i at time t, Δt is the time step, P i (fault) is the failure probability of device i;

[0098] S62, reasonable energy allocation: Dynamically adjust the energy allocation of each device according to the operating status and load demand of the device, expressed as:

[0099] E i (t+1)=Ei (t)+α·(D i (t)-E i (t));

[0100] Among them, E i (t) is the energy allocation of equipment i at time t, α is the adjustment coefficient, D i (t) is the load demand of equipment i;

[0101] S63, production process optimization and adjustment: According to the fault prediction and operation scheduling results, optimize the production process and adjust the production parameters and processes, which is expressed as:

[0102] P j (t+1)=P j (t)+β·(O j (t)-P j (t));

[0103] Among them, P j (t) is the value of production process parameter j at time t, β is the adjustment coefficient, O j (t) is the optimization target value.

[0104] The smart energy management system based on the industrial Internet is used to implement the above-mentioned smart energy management method based on the industrial Internet, and includes the following modules:

[0105] Data acquisition module: Through multimodal sensors deployed in industrial production equipment and environments, real-time energy consumption data and environmental data in the production process are collected, including electricity, heat, gas fuel, temperature, humidity and air pressure;

[0106] Data fusion module: The collected energy consumption data and environmental data are processed by data fusion technology to integrate multi-source data;

[0107] Distributed energy optimization module: Based on the results of data fusion, edge computing technology is used to optimize the energy consumption of each node in real time on distributed nodes;

[0108] Intelligent load balancing optimization module: Based on the optimized energy consumption, dynamically adjust the operating load of each industrial production equipment to achieve optimal distribution of energy load;

[0109] Multi-level fault prediction module: Combined with the results of energy load optimization allocation, it predicts faults of industrial production equipment, warns of potential faults in advance, and formulates corresponding maintenance plans;

[0110] Control strategy generation module: Combined with the fault prediction results of industrial production equipment, it generates adaptive control strategies, including equipment operation scheduling, reasonable energy allocation and production process optimization and adjustment.

[0111] Beneficial effects of the present invention:

[0112] The present invention uses a data acquisition module to acquire multimodal data such as electric energy, thermal energy, gas fuel and environmental parameters in the industrial production process in real time to ensure the integrity and accuracy of the data. The data fusion module uses a weighted average fusion algorithm and the Pearson correlation coefficient to perform multi-source data fusion and correlation analysis, which significantly improves the consistency of multi-source data and the effect of feature extraction, and provides reliable data support for subsequent energy optimization management and fault prediction.

[0113] The present invention optimizes the energy consumption of each distributed node in real time through edge computing technology, and uses a linear programming algorithm to dynamically adjust the energy allocation of each node under the premise of meeting equipment operation constraints and production needs, effectively reducing data transmission delays and improving the response speed and efficiency of energy management. The intelligent load balancing optimization module uses a Gaussian process regression model for load prediction and realizes dynamic load balancing through a particle swarm optimization algorithm, ensuring that the optimization result not only has the lowest energy consumption, but also has a uniform load distribution, avoiding local overload and energy waste, and significantly improving the overall energy utilization efficiency.

[0114] The present invention extracts fault-related features from energy consumption data and environmental data through a convolutional neural network model, and performs fault classification and prediction, warns of potential faults in advance, formulates corresponding maintenance plans, and ensures stable operation of the system. The control strategy generation module combines the fault prediction results to generate an adaptive control strategy, including equipment operation scheduling, reasonable energy allocation, and production process optimization and adjustment, dynamically adjusts the operating status and process parameters of the equipment, reduces the probability of failure, improves production efficiency, and ensures the continuity and reliability of industrial production. BRIEF DESCRIPTION OF THE DRAWINGS

[0115] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0116] Figure 1 A schematic diagram of a management method flow chart of an embodiment of the present invention;

[0117] Figure 2 Schematic diagram of system function modules according to an embodiment of the present invention. DETAILED DESCRIPTION

[0118] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0119] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0120] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0121] like Figure 1 As shown in the figure, the smart energy management method based on the industrial Internet includes the following steps:

[0122] S1, data collection: through the multimodal sensors deployed in industrial production equipment and environment, real-time collection of energy consumption data and environmental data in the production process, including electricity, heat, gas fuel, temperature, humidity, and air pressure;

[0123] S2, data fusion: The collected energy consumption data and environmental data are processed by data fusion technology to improve the integrity and accuracy of the data;

[0124] S3, Distributed Energy Optimization: Based on the results of data fusion, edge computing technology is used to optimize the energy consumption of each node in real time on distributed nodes, reduce data transmission delays, and improve the response speed of energy management;

[0125] S4, intelligent load balancing optimization: Based on the optimized energy consumption, dynamically adjust the operating load of each industrial production equipment to achieve optimal distribution of energy load, avoid local overload and energy waste, and improve overall energy utilization efficiency, including:

[0126] S41, data monitoring: real-time monitoring of energy consumption data of each industrial production equipment after optimization;

[0127] S42, load forecasting: Based on the energy consumption data of historical monitoring and real-time monitoring, an intelligent load forecasting model is used to predict the load demand of each industrial production equipment;

[0128] S43, dynamic adjustment: according to the load demand prediction results, the load balancing optimization algorithm is used to dynamically adjust the operating load of each industrial production equipment;

[0129] S5, multi-level fault prediction: Combined with the results of the optimal allocation of energy loads, fault prediction is performed on industrial production equipment, potential faults are warned in advance, and corresponding maintenance plans are formulated to ensure stable operation of the system;

[0130] S6, control strategy generation: Combined with the fault prediction results of industrial production equipment, adaptive control strategies are generated, including equipment operation scheduling, energy rational allocation, and production process optimization and adjustment;

[0131] Through the above steps, the response speed and accuracy of energy management can be comprehensively improved, energy distribution can be optimized, potential failures can be prevented, and stable operation of the system can be ensured, thereby significantly improving the overall energy utilization efficiency and economic benefits.

[0132] Data collection in S1 includes:

[0133] S11, power data collection: Deploy power sensors on industrial production equipment to collect real-time power data consumed during equipment operation;

[0134] S12, thermal energy data collection: Deploy thermal energy sensors in the heating units of industrial production equipment to monitor and collect thermal energy data generated and consumed during equipment operation in real time;

[0135] S13, gas fuel data collection: deploy gas fuel sensors in industrial production environments to collect real-time data on gas fuel consumed by gas boilers or gas equipment;

[0136] S14, environmental temperature and humidity data collection: Deploy temperature sensors and humidity sensors in industrial production environments to monitor and collect environmental temperature and humidity data in real time;

[0137] S15, air pressure data collection: deploy air pressure sensors in industrial production environments to monitor and collect ambient air pressure data in real time, helping to comprehensively evaluate the impact of the environment on energy consumption;

[0138] Through the above content, it is possible to comprehensively and accurately monitor the energy consumption and environmental changes in the production process, ensure the integrity and accuracy of the data, thereby providing reliable data support for subsequent energy optimization management and fault prediction, and significantly improving the efficiency and effectiveness of overall energy management.

[0139] Data fusion in S2 includes:

[0140] S21, data preprocessing: preliminary processing of the collected energy consumption data and environmental data, including data cleaning, denoising and normalization, to ensure data consistency and availability;

[0141] S22, fused data: The weighted average fusion algorithm is used to fuse multi-source data, which is expressed as:

[0142]

[0143] Among them, d i represents the observation value of the i-th data source, w i represents the weight of the i-th data source and satisfies

[0144] S23, data association analysis: Combined with the actual situation of the production process, the Pearson correlation coefficient is calculated to perform association analysis on the fused data, extract key features and patterns, and provide data support for subsequent energy optimization and fault prediction, which is expressed as:

[0145]

[0146] Where r is the Pearson correlation coefficient, n is the number of data points, x and y are the data values ​​from two different data sources, ∑xy is the sum of the products of x and y, ∑x and ∑y are the sum of x and y, respectively, ∑x 2 and ∑y 2 are the sum of the squares of x and y respectively;

[0147] Through the above content, the integrity and accuracy of multi-source data can be significantly improved, the consistency between different data sources can be ensured, key features and patterns can be extracted, and reliable data support can be provided for subsequent energy optimization management and fault prediction, thereby improving the effectiveness and efficiency of the overall energy management system.

[0148] Distributed energy optimization in S3 includes:

[0149] S31, Edge Node Deployment: Deploy edge computing devices on various distributed nodes of industrial production to process and optimize local energy consumption data in real time;

[0150] S32, data input: input the results of data fusion to each edge node, including consumption data of electricity, heat energy, gas fuel and environmental data;

[0151] S33, optimization target setting: set the optimization target function to minimize energy consumption while meeting production requirements and equipment operation constraints. The optimization target function F is expressed as:

[0152]

[0153] Among them, E i represents the energy consumption of the i-th node, n is the number of nodes;

[0154] S34, Constraints: Considering the production requirements of each node and the equipment operation restrictions, the constraint is set as the maximum power P of the equipment. max and minimum power P min , expressed as:

[0155] P min,i ≤P i ≤P max,i ;

[0156] Among them, P i is the actual power of the i-th node;

[0157] S35, real-time optimization calculation: Use edge computing devices to perform real-time optimization calculations, use linear programming algorithms to solve the optimization objective function, and dynamically adjust the energy consumption distribution of each node under the premise of meeting the constraints, expressed as:

[0158]

[0159] Among them, a i and b i is the optimization coefficient;

[0160] S36, result application: applying the results of the optimization calculation to each node, realizing real-time optimization of the energy consumption of each node, reducing the overall energy consumption, and improving energy utilization efficiency;

[0161] Through the above steps, energy optimization of distributed nodes can be achieved, and edge computing technology can be used to process and optimize energy consumption data in real time at each node to ensure the timeliness and accuracy of the optimization process, effectively improving the efficiency and response speed of the overall energy management system.

[0162] Data monitoring in S41 includes:

[0163] S411, Sensor deployment: Deploy various types of sensors on various industrial production equipment, including power sensors, thermal energy sensors, and gas fuel sensors, to monitor the energy consumption of the equipment in real time after optimization;

[0164] S412, data collection frequency setting: set the collection frequency of electric energy, thermal energy and gas fuel sensors according to the operating characteristics and production requirements of industrial production equipment to ensure the timeliness and accuracy of data collection;

[0165] The data collection frequency setting specifically includes:

[0166] Operation characteristics: For devices with large operation fluctuations, set a high frequency of once per second to collect data; for devices with relatively stable operation, set a medium frequency of once per minute to collect data; for devices with slow changes, set a low frequency of once per hour to collect data;

[0167] Production requirements: According to the accuracy requirements for energy consumption monitoring in the production process, set an appropriate collection frequency to ensure the needs of process control and optimization management. For example, set a high frequency of once per second for key process links and a medium frequency of once per minute for general process links;

[0168] Through the above steps, real-time monitoring of the optimized energy consumption data of each industrial production equipment can be achieved to ensure the timeliness, accuracy and effectiveness of the data, and provide reliable data support for subsequent optimization management and fault prediction.

[0169] The intelligent load forecasting model in S42 adopts the Gaussian process regression (GPR) model, which includes:

[0170] S421, define input variables and output variables: Input variable X t Contains historical energy consumption data and environmental data, output variable y t is the actual energy consumption data, expressed as:

[0171] X t =[E t-1 ,E t-2 ,...,E t-n ,θ t ];

[0172] Among them, E t-1 、E t-2 、E t-n Respectively represent the historical energy consumption data at time t-1, t-2, and tn, θ t represents the environmental data at time t;

[0173] S422, define an adaptive kernel function: Based on the characteristics of different input variables, design an adaptive kernel function k, expressed as:

[0174]

[0175] Among them, Xi and X j is the input variable, is the signal variance, l d is the length scale of the dth dimension, reflecting the adaptive characteristics, is the noise variance, δ ij is the Kronecker delta function, which takes the value 1 when i = j and 0 otherwise;

[0176] S423, constructing a covariance matrix: constructing a covariance matrix K according to the adaptive kernel function, expressed as:

[0177]

[0178] Among them, K(X,X′) is the covariance matrix, which describes the correlation between different input variables;

[0179] S424, predictive distribution: predict new energy consumption data y based on the principle of Gaussian process regression * The distribution of is expressed as:

[0180]

[0181] cov(f * )=K(X * ,X * )-K(X * ,X)K(X,X) -1 K(X,X * );

[0182] in, is the predicted mean, which represents the predicted value of the new energy consumption data, cov(f * ) is the prediction covariance, which indicates the confidence interval of the prediction value and reflects the uncertainty of the prediction result;

[0183] S425, Dynamically update model: Dynamically update the Gaussian process regression model according to real-time data, expressed as:

[0184]

[0185] Among them, θ represents the model parameters, α t is the adaptive learning rate at time t, is the gradient of the loss function L with respect to the parameter θ;

[0186] Through the above content, multi-variable input, adaptive kernel function and dynamic update mechanism are introduced, so that it can comprehensively consider historical energy consumption data, environmental data and operating status data, thereby improving the accuracy and real-time performance of load forecasting. The adaptive kernel function enhances the flexibility of the model under different input variables, and the dynamic update mechanism ensures that the model can reflect the latest production environment changes in a timely manner, so that the GPR model can more accurately predict the load demand of industrial production equipment, provide reliable data support for energy consumption optimization, and significantly improve the efficiency and response speed of the energy management system.

[0187] The load balancing optimization algorithm in S43 adopts the particle swarm optimization algorithm (PSO), which includes:

[0188] S431, Initialization: Initialize the particle swarm, including the position and velocity of the particles, set the swarm size N, the initial inertia weight w, and the learning factor c 1 and c 2 , maximum number of iterations T, current number of iterations t, expressed as:

[0189] Particle position initialization: x i (0)~U(x min ,x max );

[0190] Particle velocity initialization: v i (0)~U(v min ,v max );

[0191] Among them, U represents uniform distribution, x min and x max are the minimum and maximum values ​​of the position, respectively, v min and v max are the minimum and maximum values ​​of the speed respectively;

[0192] S432, velocity update: based on the particle's own historical best position p i (t) and the global optimal position g(t), introduce the adaptive inertia weight w(t), and update the particle velocity, which is expressed as:

[0193]

[0194] v i (t+1)=w(t)·v i (t)+c 1 ·r 1 ·(p i (t)-x i (t))+c 2 ·r 2 ·(g(t)-xi (t));

[0195] Among them, w max is the initial inertia weight, w min is the final inertia weight, T is the maximum number of iterations, t is the current number of iterations, r 1 and r 2 is a random number with a value range of [0,1], c 1 and c 2 is the learning factor, which represents the influence of the particle's own experience and the group's experience on the speed update, and its value is 2, x i (t) is the position of particle i at time t, representing the operating load of the i-th device, v i (t+1) is the speed of particle i at time t+1, indicating the updated load adjustment speed, v i (t) is the speed of particle i at time t, indicating the speed of load adjustment;

[0196] S433, position update: Update the particle's position according to the new velocity, expressed as:

[0197] x i (t+1)=x i (t)+v i (t+1);

[0198] Among them, x i (t+1) is the position of particle i at time t+1, indicating the adjusted operating load of the i-th device;

[0199] S434, fitness evaluation: calculate the fitness value f(x i (t+1)), a multi-objective fitness function is used to comprehensively consider energy consumption and load balance, which is expressed as:

[0200] f(x i )=α·E i +β·B i ;

[0201]

[0202] Among them, E i is the energy consumption of the ith device, B i is the load balancing metric, P j is the actual power of the jth device, α and β are weight coefficients, which control the weights of energy consumption and load balance;

[0203] Update the best historical position of the particle:

[0204] Update the global best position: g(t+1)=min i f(p i (t+1));

[0205] S435, iteration: repeat speed update to fitness evaluation until the stop condition is reached (such as the maximum number of iterations T or the fitness function converges);

[0206] Through the above content, the particle swarm is initialized on the industrial production equipment, including the initial load and adjustment speed of each device. Through the adaptive inertia weight, the algorithm can perform global search in the early stage of iteration and local search in the later stage of iteration to improve the optimization efficiency. The speed is adjusted according to the new load, and the operating load of each device is updated. A multi-objective fitness function is adopted to comprehensively consider energy consumption and load balance to ensure that the optimization result not only has the lowest energy consumption, but also has an evenly distributed load to avoid local overload. The optimization process is repeated until the optimal load distribution plan is found, which can better adapt to the needs of intelligent load balancing optimization, dynamically adjust the operating load of each industrial production equipment, achieve energy consumption optimization and load balance, and improve overall energy utilization efficiency.

[0207] The multi-level fault prediction in S5 includes:

[0208] S51, data collection and preprocessing: Collect energy consumption data and environmental data from industrial production equipment, and perform data cleaning, denoising and normalization;

[0209] S52, feature extraction: A convolutional neural network (CNN) model is used to extract fault-related features from the preprocessed energy consumption data and environmental data, expressed as:

[0210] Convolutional Layer:

[0211] Where X is the input data, W is the convolution kernel weight, b is the bias, σ is the activation function, F is the convolution layer output feature map, i, j are the spatial position indexes, k is the feature map index, m, n, p, M, N, P are the convolution kernel sizes;

[0212] Pooling layer:

[0213] Among them, Q is the output feature map of the pooling layer, a, b, A, B are the pooling window sizes;

[0214] S53, fault classification and prediction: Use convolutional neural network (CNN) model for fault classification and prediction, expressed as:

[0215] Flatten a vector: h = flatten(Q);

[0216] First fully connected layer: y 1 =σ(W1 ·h+b 1 );

[0217] Second fully connected layer (output layer):

[0218] Among them, W 1 is the weight matrix, b 1 is the bias vector, y 1 is the output vector, W 2 is the weight matrix, b 2 is the bias vector, is the output probability vector;

[0219] S54, fault probability calculation: The output results of the convolutional neural network (CNN) model are used to calculate the probability of each fault occurring, expressed as:

[0220]

[0221] where z is the eigenvector, is the probability vector of failure occurrence;

[0222] S55, maintenance plan formulation: Based on the fault prediction results and the probability of fault occurrence, early warning of potential faults is provided, and corresponding maintenance plans are formulated;

[0223] The maintenance plan specifically includes:

[0224] Regular inspection and maintenance: Regular inspection and maintenance should be carried out according to the use of the equipment;

[0225] Preventive maintenance: Based on the fault prediction results, equipment that may have problems can be maintained in advance;

[0226] Troubleshooting and repair: When a fault is detected, quickly troubleshoot and repair it;

[0227] Spare parts management: ensuring the inventory of commonly used spare parts so that faulty parts can be replaced quickly;

[0228] Equipment upgrade and update: Regularly evaluate the performance of equipment and upgrade and update when necessary;

[0229] Training and records: Provide training to maintenance personnel and keep detailed records of the circumstances and results of each maintenance.

[0230] Through the above content, the accuracy of fault prediction can be improved, problems can be discovered and solved in time, equipment downtime can be reduced, maintenance costs can be reduced, production efficiency can be improved, and the continuity and reliability of industrial production can be ensured.

[0231] The control strategy generation in S6 includes:

[0232] S61, equipment operation scheduling: According to the equipment failure prediction results, the equipment operation time and load are adjusted to reduce the risk of failure, which is expressed as:

[0233] S i (t+1)=S i (t)+Δt·(1-P i (fault));

[0234] Among them, S i (t) is the scheduling state of device i at time t, Δt is the time step, P i (fault) is the failure probability of device i;

[0235] S62, reasonable energy allocation: According to the operating status and load demand of the equipment, dynamically adjust the energy allocation of each device to improve the overall energy utilization efficiency, expressed as:

[0236] E i (t+1)=E i (t)+α·(D i (t)-E i (t));

[0237] Among them, E i (t) is the energy allocation of equipment i at time t, α is the adjustment coefficient, D i (t) is the load demand of equipment i;

[0238] S63, production process optimization and adjustment: According to the fault prediction and operation scheduling results, optimize the production process, adjust the production parameters and processes to reduce the risk of failure and improve production efficiency, expressed as:

[0239] P j (t+1)=P j (t)+β·(O j (t)-P j (t));

[0240] Among them, P j (t) is the value of production process parameter j at time t, β is the adjustment coefficient, O j (t) is the optimization target value;

[0241] Through the above content, the operating status of the equipment can be adjusted in time, the probability of failure can be reduced, energy distribution can be optimized to reduce waste, and production efficiency can be improved by adjusting the production process. This dynamic and intelligent control strategy not only reduces maintenance costs and equipment downtime, but also improves the reliability and economy of the overall production process.

[0242] like Figure 2As shown, the smart energy management system based on the industrial Internet is used to implement the above-mentioned smart energy management method based on the industrial Internet, and includes the following modules:

[0243] Data acquisition module: Through multimodal sensors deployed in industrial production equipment and environments, real-time energy consumption data and environmental data in the production process are collected, including electricity, heat, gas fuel, temperature, humidity and air pressure;

[0244] Data fusion module: The collected energy consumption data and environmental data are processed by data fusion technology to integrate multi-source data;

[0245] Distributed energy optimization module: Based on the results of data fusion, edge computing technology is used to optimize the energy consumption of each node in real time on distributed nodes;

[0246] Intelligent load balancing optimization module: Based on the optimized energy consumption, dynamically adjust the operating load of each industrial production equipment to achieve optimal distribution of energy load;

[0247] Multi-level fault prediction module: Combined with the results of energy load optimization allocation, it predicts faults of industrial production equipment, warns of potential faults in advance, and formulates corresponding maintenance plans;

[0248] Control strategy generation module: Combined with the fault prediction results of industrial production equipment, it generates adaptive control strategies, including equipment operation scheduling, reasonable energy allocation and production process optimization and adjustment.

[0249] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0250] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. The smart energy management method based on the Industrial Internet is characterized by: The following steps are involved: S1, data collection: through the multimodal sensors deployed in industrial production equipment and environment, real-time collection of energy consumption data and environmental data in the production process, including electricity, heat, gas fuel, temperature, humidity, and air pressure; S2, data fusion: the collected energy consumption data and environmental data are fused by using data fusion technology to perform multi-source data fusion processing; S3, distributed energy optimization: Based on the results of data fusion, edge computing technology is used to optimize the energy consumption of each node in real time on distributed nodes; S4, intelligent load balancing optimization: Based on the optimized energy consumption, dynamically adjust the operating load of each industrial production equipment to achieve optimal distribution of energy load, including: S41, data monitoring: real-time monitoring of energy consumption data of each industrial production equipment after optimization; S42, load forecasting: Based on the energy consumption data of historical monitoring and real-time monitoring, an intelligent load forecasting model is used to predict the load demand of each industrial production equipment; S43, dynamic adjustment: according to the load demand prediction results, the load balancing optimization algorithm is used to dynamically adjust the operating load of each industrial production equipment; S5, multi-level fault prediction: Combined with the results of the optimal allocation of energy loads, fault prediction is performed on industrial production equipment, potential faults are warned in advance, and corresponding maintenance plans are formulated; S6, control strategy generation: Combined with the fault prediction results of industrial production equipment, generate adaptive control strategies, including equipment operation scheduling, reasonable energy allocation, and production process optimization and adjustment.

2. The smart energy management method based on industrial Internet according to claim 1 is characterized in that: The data collection in S1 includes: S11, power data collection: Deploy power sensors on industrial production equipment to collect real-time power data consumed during equipment operation; S12, thermal energy data collection: Deploy thermal energy sensors in the heating units of industrial production equipment to monitor and collect thermal energy data generated and consumed during equipment operation in real time; S13, gas fuel data collection: deploy gas fuel sensors in industrial production environments to collect real-time data on gas fuel consumed by gas boilers or gas equipment; S14, environmental temperature and humidity data collection: Deploy temperature sensors and humidity sensors in industrial production environments to monitor and collect environmental temperature and humidity data in real time; S15, air pressure data collection: deploy air pressure sensors in industrial production environments to monitor and collect ambient air pressure data in real time.

3. The smart energy management method based on industrial Internet according to claim 1 is characterized in that: The data fusion in S2 includes: S21, data preprocessing: preliminary processing of the collected energy consumption data and environmental data, including data cleaning, denoising and normalization; S22, fusion data: use weighted average fusion algorithm to fuse multi-source data; S23, data association analysis: Based on the actual situation of the production process, the fused data is analyzed by calculating the Pearson correlation coefficient to extract key features and patterns.

4. The smart energy management method based on industrial Internet according to claim 1 is characterized in that: The distributed energy optimization in S3 includes: S31, Edge Node Deployment: Deploy edge computing devices on various distributed nodes of industrial production to process and optimize local energy consumption data in real time; S32, data input: input the results of data fusion to each edge node, including consumption data of electricity, heat energy, gas fuel and environmental data; S33, optimization target setting: set the optimization target function to minimize energy consumption while meeting production requirements and equipment operation constraints. The optimization target function F is expressed as: Among them, E i represents the energy consumption of the i-th node, n is the number of nodes; S34, Constraints: Considering the production requirements of each node and the equipment operation restrictions, the constraint is set as the maximum power P of the equipment. max and minimum power P min , expressed as: P min,i ≤P i ≤P max,i ; Among them, P i is the actual power of the i-th node; S35, real-time optimization calculation: Use edge computing devices to perform real-time optimization calculations, use linear programming algorithms to solve the optimization objective function, and dynamically adjust the energy consumption distribution of each node under the premise of meeting the constraints, expressed as: Among them, a i and b i is the optimization coefficient; S36, result application: applying the result of the optimization calculation to each node to achieve real-time optimization of the energy consumption of each node.

5. The smart energy management method based on industrial Internet according to claim 4 is characterized in that: The data monitoring in S41 includes: S411, Sensor deployment: Deploy various types of sensors on various industrial production equipment, including power sensors, thermal energy sensors, and gas fuel sensors, to monitor the energy consumption of the equipment in real time after optimization; S412, data collection frequency setting: setting the collection frequency of the electric energy, thermal energy and gas fuel sensors according to the operating characteristics and production requirements of the industrial production equipment.

6. The smart energy management method based on industrial Internet according to claim 5 is characterized in that: The intelligent load forecasting model in S42 adopts a Gaussian process regression model, and the Gaussian process regression model includes: S421, define input variables and output variables: Input variable X t Contains historical energy consumption data and environmental data, output variable y t is the actual energy consumption data, expressed as: X t =[And t-1 ,AND t-2 ,...,AND t-n ,θ t ]; Among them, E t-1 、E t-2 、E t-n Respectively represent the historical energy consumption data at time t-1, t-2, and tn, θ t represents the environmental data at time t; S422, define an adaptive kernel function: Based on the characteristics of different input variables, design an adaptive kernel function k, expressed as: Among them, X i and X j is the input variable, is the signal variance, l d is the length scale of the dth dimension, is the noise variance, δ ij is the Kronecker delta function; S423, constructing a covariance matrix: constructing a covariance matrix K according to the adaptive kernel function, expressed as: Where K(X,X′) is the covariance matrix; S424, predict distribution: According to the principle of Gaussian process regression, predict the distribution of new energy consumption data y*, expressed as: those(f * )=K(X * ,X * )-K(X * ,X)K(X,X) -1 K(X,X * ); in, is the predicted mean, cov(f * ) is the prediction covariance; S425, Dynamically update model: Dynamically update the Gaussian process regression model according to real-time data, expressed as: Among them, θ represents the model parameters, α t is the adaptive learning rate at time t, is the gradient of the loss function L with respect to the parameter θ.

7. The smart energy management method based on industrial Internet according to claim 6 is characterized in that: The load balancing optimization algorithm in S43 adopts a particle swarm optimization algorithm, and the particle swarm optimization algorithm includes: S431, initialization: Initialize the particle swarm, including the position and velocity of the particles, set the swarm size N, the initial inertia weight w, the learning factors c1 and c2, the maximum number of iterations T, and the current number of iterations t, expressed as: Particle position initialization: x i (0)~U(x min ,x max ); Particle velocity initialization: v i (0)~U(v min ,v max ); Among them, U represents uniform distribution, x min and x max are the minimum and maximum values ​​of the position, respectively, v min and v max are the minimum and maximum values ​​of the speed respectively; S432, velocity update: based on the particle's own historical best position p i (t) and the global optimal position g(t), introduce the adaptive inertia weight w(t), and update the particle velocity, which is expressed as: v i (t+1)=w(t)·v i (t)+c1·r1·(p i (t)-x i (t))+c2·r2·(g(t)-x i (t)); Among them, w max is the initial inertia weight, w min is the final inertia weight, T is the maximum number of iterations, t is the current number of iterations, r1 and r2 are random numbers, c1 and c2 are learning factors, and x i (t) is the position of particle i at time t, v i (t+1) is the velocity of particle i at time t+1, v i (t) is the velocity of particle i at time t; S433, position update: Update the particle's position according to the new velocity, expressed as: x i (t+1)=x i (t)+v i (t+1); Among them, x i (t+1) is the position of particle i at time t+1; S434, fitness evaluation: calculate the fitness value f(x i (t+1)), a multi-objective fitness function is used to comprehensively consider energy consumption and load balance, which is expressed as: f(x i )=α·E i +β·B i ; Among them, E i is the energy consumption of the ith device, B i is the load balancing measure, P j is the actual power of the jth device, α and β are weight coefficients; Update the best historical position of the particle: Update the global best position: g(t+1)=min i f(p i (t+1)); S435, Iteration: Repeat speed update to fitness evaluation until the stop condition is reached.

8. The smart energy management method based on industrial Internet according to claim 7 is characterized in that: The multi-level fault prediction in S5 includes: S51, data collection and preprocessing: Collect energy consumption data and environmental data from industrial production equipment, and perform data cleaning, denoising and normalization; S52, feature extraction: A convolutional neural network model is used to extract fault-related features from the preprocessed energy consumption data and environmental data, expressed as: Convolutional Layer: Where X is the input data, W is the convolution kernel weight, b is the bias, σ is the activation function, F is the convolution layer output feature map, i, j are the spatial position indexes, k is the feature map index, m, n, p, M, N, P are the convolution kernel sizes; Pooling layer: Among them, Q is the output feature map of the pooling layer, a, b, A, B are the pooling window sizes; S53, Fault Classification and Prediction: Fault classification and prediction are performed using a convolutional neural network model, expressed as: Flatten a vector: h = flatten(Q); First fully connected layer: y1 = σ(W1 h + b1); Second fully connected layer: Among them, W1 is the weight matrix, b1 is the bias vector, y1 is the output vector, W2 is the weight matrix, b2 is the bias vector, is the output probability vector; S54, fault probability calculation: Use the output results of the convolutional neural network model to calculate the probability of each fault occurring, expressed as: where z is the eigenvector, is the probability vector of failure occurrence; S55, maintenance plan formulation: Based on the fault prediction results and the probability of fault occurrence, potential faults are warned in advance and corresponding maintenance plans are formulated.

9. The smart energy management method based on industrial Internet according to claim 8 is characterized in that: The control strategy generation in S6 includes: S61, equipment operation scheduling: according to the equipment fault prediction results, adjust the equipment operation time and load, expressed as: S i (t+1)=S i (t)+Δt·(1-P i (fault)); Among them, S i (t) is the scheduling state of device i at time t, Δt is the time step, P i (fault) is the failure probability of device i; S62, reasonable energy allocation: Dynamically adjust the energy allocation of each device according to the operating status and load demand of the device, expressed as: E i (t+1)=E i (t)+α·(D i (t)-E i (t)); Among them, E i (t) is the energy allocation of equipment i at time t, α is the adjustment coefficient, D i (t) is the load demand of equipment i; S63, production process optimization and adjustment: according to the fault prediction and operation scheduling results, optimize the production process and adjust the production parameters and processes, expressed as: P j (t+1)=P j (t)+β·(O j (t)-P j (t)); Among them, P j (t) is the value of production process parameter j at time t, β is the adjustment coefficient, O j (t) is the optimization target value.

10. A smart energy management system based on the industrial Internet, used to implement the smart energy management method based on the industrial Internet as described in any one of claims 1 to 9, characterized in that: Includes the following modules: Data acquisition module: Through multimodal sensors deployed in industrial production equipment and environments, real-time energy consumption data and environmental data in the production process are collected, including electricity, heat, gas fuel, temperature, humidity and air pressure; Data fusion module: The collected energy consumption data and environmental data are processed by data fusion technology to integrate multi-source data; Distributed energy optimization module: Based on the results of data fusion, edge computing technology is used to optimize the energy consumption of each node in real time on distributed nodes; Intelligent load balancing optimization module: Based on the optimized energy consumption, dynamically adjust the operating load of each industrial production equipment to achieve optimal distribution of energy load; Multi-level fault prediction module: Combined with the results of energy load optimization allocation, it predicts faults of industrial production equipment, warns of potential faults in advance, and formulates corresponding maintenance plans; Control strategy generation module: Combined with the fault prediction results of industrial production equipment, it generates adaptive control strategies, including equipment operation scheduling, reasonable energy allocation and production process optimization and adjustment.

Images