Intelligent monitoring system for multi-source data fusion of intelligent biomass power plant

Abstract

The invention discloses an intelligent monitoring system for multi-source data fusion of a biomass intelligent power plant, relates to the technical field of intelligent monitoring, and solves the problems proposed in the background technology. The biomass multi-source data acquisition module is used for acquiring biomass fuel characteristic data, equipment operation data, environment-friendly emission data and raw material supply chain data in real time; the data preprocessing and order reduction module is used for carrying out denoising, normalization and coupling order reduction processing on the biomass fuel characteristic data and the equipment operation data; the multi-source data dynamic fusion module is used for carrying out weighted fusion on the multi-source data according to the fuel heat value, the water content and the environmental protection weight; the biomass combustion optimization diagnosis module is used for constructing a deep learning model based on the fusion data and outputting combustion efficiency optimization parameters and a fault diagnosis result; according to the method, multi-dimensional data of fuel characteristics, equipment operation, environmental protection emission and the like are integrated, through links of preprocessing, order reduction, fusion and the like, an information island phenomenon is eliminated, and scientificity and accuracy of decision making are improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent monitoring technology, and in particular to an intelligent monitoring system for multi-source data fusion of a biomass smart power plant. Background Art

[0002] As clean energy gradually becomes an important direction for the transformation of energy structure, biomass energy, as an important component of renewable energy, has the characteristics of a wide range of raw materials and has been widely used in cogeneration, power plant heating, rural energy supply and other fields. Biomass power plants generate electricity by burning raw materials such as straw, wood chips, and forestry residues. Their operating efficiency and environmental protection indicators are affected by the volatility of fuel properties and the complexity of equipment operating conditions. Therefore, higher requirements are placed on the intelligence and precision of the monitoring system.

[0003] Publication No. CN109709839A discloses a multi-level real-time monitoring method for biomass power plants, including the following steps: collecting and aggregating real-time data from all equipment measurement points and calculation point data; pushing data in real time, converting OPC protocol data into UDP protocol data; sending the data in real time to a mirrored sending server via the UDP protocol; receiving the mirrored sent data and forwarding it in real time to the mirrored receiving server; converting the real-time collected monitoring point data into target data; and performing trend analysis on the collected data to achieve the goal of monitoring the production status and trends of each biomass power plant. This method integrates production operation information and production management information, streamlines business operations, and provides scientific decision support, helping users effectively improve the efficiency of biomass power plants.

[0004] The system only realizes the aggregation and classification display of data, lacks processing methods such as coupled order reduction and weighted fusion of multi-source heterogeneous data, and cannot extract deep correlation features between cross-source data. The system does not introduce deep learning or intelligent diagnostic models, and cannot adaptively model and predict complex combustion states and operational faults. It still relies on manual monitoring and experience judgment. To this end, we proposed an intelligent monitoring system for multi-source data fusion of biomass smart power plants. Summary of the Invention

[0005] To this end, the present application provides an intelligent monitoring system for multi-source data fusion of biomass smart power plants to solve the problem that the existing technology only realizes data aggregation and classification display, lacks processing methods such as coupling reduction and weighted fusion of multi-source heterogeneous data, and cannot extract deep correlation features between cross-source data.

[0006] In order to achieve the above objectives, this application provides the following technical solutions:

[0007] Firstly, the system includes:

[0008] Biomass multi-source data acquisition module: used to collect biomass fuel characteristic data, equipment operation data, environmental emission data and raw material supply chain data in real time;

[0009] Data preprocessing and order reduction module: performs denoising, normalization and coupled order reduction processing on the biomass fuel characteristic data and equipment operation data;

[0010] Multi-source data dynamic fusion module: weighted fusion of multi-source data based on fuel calorific value, moisture content and environmental protection weight;

[0011] Biomass combustion optimization and diagnosis module: Builds a deep learning model based on fused data to output combustion efficiency optimization parameters and fault diagnosis results;

[0012] Intelligent feedback control module: dynamically adjusts combustion process parameters, environmental control strategies and raw material supply chain management instructions based on diagnostic results,

[0013] Optionally, the biomass multi-source data acquisition module includes:

[0014] Fuel property detection unit: equipped with a near-infrared spectrometer to detect the moisture content, ash content and calorific value of biomass raw materials in real time;

[0015] Equipment operation sensor unit: collects boiler temperature, combustion chamber pressure, flue gas oxygen content and vibration data;

[0016] Environmental monitoring unit: detects particulate matter concentration, NOx concentration and CO2 emissions;

[0017] Supply chain management unit: obtain the real-time location of raw material transportation, storage humidity and remaining inventory days.

[0018] Optionally, the method for the data preprocessing and order reduction module to perform denoising, normalization, and coupled order reduction processing specifically includes:

[0019] A1. Dynamic compensation of calorific value: Dynamically correct the original calorific value according to the moisture content to generate a compensated calorific value, where the compensation coefficient is related to the biomass fuel type;

[0020] A2. Segment-wise normalization of environmental data: nonlinear compression processing of pollutant concentrations exceeding the standard;

[0021] A3. Coupling matrix reduction: Perform joint singular value decomposition on the fuel characteristic matrix and the equipment operation matrix to retain the principal component characteristics related to combustion efficiency.

[0022] Optionally, the weight distribution of the multi-source data dynamic fusion module satisfies: the weight is comprehensively calculated by the fuel calorific value stability coefficient, the pollutant emission priority coefficient and the biomass type adjustment factor, wherein the adjustment factor is set differently according to the type of straw, sawdust and biogas.

[0023] Optionally, the biomass combustion optimization diagnosis module includes:

[0024] Dual-channel deep residual network: processes equipment operation data and fuel characteristic data separately, and fuses them to output combustion efficiency and failure probability;

[0025] Multi-objective loss function: By jointly optimizing the combustion efficiency error and the dynamically adjusted environmental penalty term, the model is constrained to learn the emission limit.

[0026] Optionally, the specific method for the intelligent feedback control module to implement dynamic adjustment includes:

[0027] B1. Moisture content over-limit control: When the moisture content is detected to exceed the threshold, the drying equipment power and feed rate are adjusted in conjunction;

[0028] B2. Coking risk warning: Generate maintenance instructions based on the correlation model between ash content and combustion chamber temperature gradient;

[0029] B3. Dynamic environmental optimization: Automatically adjust the secondary air valve and urea injection amount according to the NOx concentration exceeding the standard.

[0030] Optionally, the power regulation of the drying equipment in the intelligent feedback control module dynamically sets the target temperature range according to the compensated calorific value calculated in real time, and realizes closed-loop control in combination with the change of the feed rate, wherein:

[0031] The upper and lower limits of the target temperature range are positively correlated with the compensated calorific value;

[0032] The feed rate is dynamically adjusted based on the deviation between the target temperature and the actual temperature. The greater the deviation, the greater the rate reduction.

[0033] Optionally, in the dual-channel deep residual network:

[0034] Fuel characteristic channel: A one-dimensional convolution structure is used to extract the time series variation characteristics of calorific value and moisture content;

[0035] Equipment operation channel: Multi-scale residual learning is performed on boiler temperature and flue gas oxygen content, and the output is fused and passed to the fully connected layer;

[0036] The multi-scale residual learning includes parallel processing of device operation data by convolution kernels of at least three different scales.

[0037] Optionally, the joint singular value decomposition in the coupling matrix reduction improves the sensitivity of combustion efficiency by introducing a device operating state weight matrix, where:

[0038] The weight matrix is ​​dynamically calculated based on the product of boiler temperature and flue gas oxygen content;

[0039] The total correlation coefficient between the number of principal components retained after order reduction and combustion efficiency is no less than 80%.

[0040] Compared with the prior art, this application has at least the following beneficial effects:

[0041] 1. The monitoring system of the present invention integrates data from multiple dimensions, including fuel characteristics, equipment operation, environmental emissions, and supply chain. Through preprocessing, order reduction, and fusion, it effectively eliminates information silos and improves the scientificity and accuracy of decision-making.

[0042] 2. The monitoring system of the present invention uses a dual-channel deep residual network model to diagnose the combustion status in real time and output optimization parameters, thereby ensuring combustion stability while improving thermal efficiency, reducing energy waste, and reducing fuel usage costs.

[0043] 3. The monitoring system of the present invention has functions such as coking risk warning and moisture content abnormality adjustment through coupling analysis of combustion status and equipment parameters, which can trigger maintenance instructions in advance, reduce the probability of sudden failures, and increase the service life of equipment.

[0044] 4. The monitoring system of the present invention provides differentiated parameter adjustment mechanisms for different types of biomass (such as straw, wood chips, and biogas) systems, making it flexible in deployment across fuel types and conducive to promotion. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] To more intuitively illustrate the prior art and the present application, exemplary drawings are provided below. It should be understood that the specific shapes and structures shown in the drawings should not generally be considered as limiting conditions for implementing the present application; for example, based on the technical concepts disclosed in this application and the exemplary drawings, those skilled in the art are capable of easily making routine adjustments or optional optimizations to the addition / reduction / attribution division, specific shapes, positional relationships, connection methods, dimensional ratios, etc. of certain units (components).

[0046] Figure 1 Schematic diagram of the working steps of the intelligent monitoring system for multi-source data fusion of biomass smart power plants of the present invention. DETAILED DESCRIPTION

[0047] The present application will be further described below in detail through specific embodiments in conjunction with the accompanying drawings.

[0048] Depend on Figure 1This paper proposes an intelligent monitoring system for biomass smart power plants that integrates multi-source data. The system includes a biomass multi-source data acquisition module, a data preprocessing and order reduction module, a multi-source data dynamic fusion module, a biomass combustion optimization and diagnosis module, and an intelligent feedback control module. Through the collaboration of these modules, the system enables full lifecycle monitoring and intelligent feedback control of fuel, equipment, emissions, and supply chain information, creating a closed-loop, efficient monitoring system.

[0049] The biomass multi-source data acquisition module includes a fuel property detection unit, an equipment operation sensor unit, an environmental monitoring unit and a supply chain management unit.

[0050] The fuel property detection unit is equipped with a near-infrared spectrometer to achieve real-time monitoring of the moisture content, ash content and calorific value of biomass raw materials;

[0051] The equipment operation sensor unit collects information such as boiler temperature, combustion chamber pressure, flue gas oxygen content, and equipment vibration for dynamic evaluation of equipment operation status;

[0052] The environmental monitoring unit is used to obtain particulate matter concentration, NOx concentration and CO3 emissions to reflect whether the emissions meet the standards;

[0053] The supply chain management unit works with GPS and warehouse sensors to obtain key indicators such as the real-time location of raw material transportation, warehouse humidity, and remaining inventory days.

[0054] The processing methods performed by the data preprocessing and order reduction module include:

[0055] A1. Dynamic compensation of calorific value: based on the original calorific value Q of the moisture content M raw Compensate and compensate the calorific value Q comp The calculation formula is as follows:

[0056] Q comp =Q raw ×(1-k m ×M), where k m The coefficients are set for different types of biomass (such as straw, wood chips, etc.).

[0057] A2. Segment-wise normalization of environmental data: nonlinear compression processing of pollutant concentrations exceeding the standard;

[0058] A3. Coupling matrix reduction processing:

[0059] Construct a joint matrix C of the fuel characteristic matrix and the equipment operation matrix, and extract the principal component features related to combustion efficiency through joint singular value decomposition (SVD):

[0060] The fuel calorific value stability coefficient reflects the calorific value fluctuation range; the emission priority coefficient takes into account the importance ranking of pollution indicators such as particulate matter and NOx; the biomass type adjustment factor is corrected according to the differences in the characteristics of different raw materials such as straw, sawdust, and biogas.

[0061] Biomass combustion optimization diagnostic module includes:

[0062] (1) A dual-channel deep residual network processes equipment operation data and fuel characteristic data separately. The equipment channel extracts dynamic parameter features such as temperature and pressure, while the fuel channel analyzes the temporal changes of moisture content and calorific value.

[0063] (2) Multi-objective loss function, which jointly optimizes combustion efficiency and emission indicators, is as follows:

[0064] L=λ1(η pred -η true ) 2 +λ2Penalty env

[0065] Where η represents the combustion efficiency, η pred is the theoretical value of combustion efficiency, η true is the true value of combustion efficiency, Penalty env is the penalty function for exceeding the standard of pollutants such as NOx and CO3, λ1 and λ2 represent the weight coefficient of the combustion efficiency error term and the weight coefficient of the environmental protection penalty term, respectively.

[0066] The intelligent feedback control module can implement the following controls based on the diagnosis results:

[0067] B1. Moisture content over-limit control: When the real-time moisture content exceeds the set threshold, the heating power of the drying equipment is automatically increased and the feed rate is reduced to avoid furnace temperature fluctuations;

[0068] B2. Coking risk warning: When the ash content and temperature gradient change trends indicate possible coking, the system automatically issues maintenance instructions;

[0069] B3. Environmental protection dynamic optimization: When it is detected that NOx exceeds the standard, the secondary air door angle and urea injection ratio are automatically adjusted to improve denitrification efficiency.

[0070] The dual-channel network further has the following characteristics:

[0071] The fuel characteristic channel uses a one-dimensional convolution structure to extract the time series characteristics of calorific value and moisture content;

[0072] The equipment operation channel uses three scales (convolution kernel sizes are 3, 5, and 7) of parallel residual blocks to extract multi-scale features, effectively enhancing the sensitivity to operation anomalies; the multi-scale convolution results are fused and input into the fully connected layer for output.

[0073] The optimized reduced-order method introduces an equipment state weight matrix to improve sensitivity to changes in combustion efficiency: the weight matrix is ​​constructed by multiplying the boiler temperature and the flue gas oxygen content;

[0074] The total correlation coefficient between the number of principal components retained after order reduction and combustion efficiency is no less than 80%.

[0075] The technical features of the above embodiments can be combined arbitrarily (as long as there is no contradiction in the combination of these technical features). In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described; these embodiments that are not explicitly written should also be considered to be within the scope of this specification.

Claims

1. An intelligent monitoring system for multi-source data fusion of biomass smart power plants, characterized by: include: Biomass multi-source data acquisition module: used to collect biomass fuel characteristic data, equipment operation data, environmental emission data and raw material supply chain data in real time; Data preprocessing and order reduction module: performs denoising, normalization and coupled order reduction processing on the biomass fuel characteristic data and equipment operation data; Multi-source data dynamic fusion module: weighted fusion of multi-source data based on fuel calorific value, moisture content and environmental protection weight; Biomass combustion optimization and diagnosis module: Builds a deep learning model based on fused data to output combustion efficiency optimization parameters and fault diagnosis results; Intelligent feedback control module: Dynamically adjusts combustion process parameters, environmental control strategies and raw material supply chain management instructions based on diagnostic results.

2. The intelligent monitoring system for multi-source data fusion of a biomass smart power plant according to claim 1 is characterized by: The biomass multi-source data acquisition module includes: Fuel property detection unit: equipped with a near-infrared spectrometer to detect the moisture content, ash content and calorific value of biomass raw materials in real time; Equipment operation sensor unit: collects boiler temperature, combustion chamber pressure, flue gas oxygen content and vibration data; Environmental monitoring unit: detects particulate matter concentration, NOx concentration and CO2 emissions; Supply chain management unit: obtain the real-time location of raw material transportation, storage humidity and remaining inventory days.

3. The intelligent monitoring system for multi-source data fusion of a biomass smart power plant according to claim 2 is characterized by: The method for performing denoising, normalization and coupled order reduction processing by the data preprocessing and order reduction module specifically includes: A1. Dynamic compensation of calorific value: Dynamically correct the original calorific value according to the moisture content to generate a compensated calorific value, where the compensation coefficient is related to the biomass fuel type; A2. Segment-wise normalization of environmental data: nonlinear compression processing of pollutant concentrations exceeding the standard; A3. Coupling matrix reduction: Perform joint singular value decomposition on the fuel characteristic matrix and the equipment operation matrix to retain the principal component characteristics related to combustion efficiency.

4. The intelligent monitoring system for multi-source data fusion of a biomass smart power plant according to claim 3 is characterized by: The weight distribution of the multi-source data dynamic fusion module satisfies: the weight is comprehensively calculated by the fuel calorific value stability coefficient, the pollutant emission priority coefficient and the biomass type adjustment factor, wherein the adjustment factor is set differently according to the types of straw, sawdust and biogas.

5. The intelligent monitoring system for multi-source data fusion of a biomass smart power plant according to claim 4 is characterized by: The biomass combustion optimization diagnosis module includes: Dual-channel deep residual network: includes an equipment operation channel and a fuel characteristic channel, which process equipment operation data and fuel characteristic data respectively, and fuse them to output combustion efficiency and failure probability; Multi-objective loss function: By jointly optimizing the combustion efficiency error and the dynamically adjusted environmental penalty term, the model is constrained to learn the emission limit.

6. The intelligent monitoring system for multi-source data fusion of a biomass smart power plant according to claim 5 is characterized by: The specific method for the intelligent feedback control module to achieve dynamic adjustment includes: B1. Moisture content over-limit control: When the moisture content is detected to exceed the threshold, the drying equipment power and feed rate are adjusted in conjunction; B2. Coking risk warning: Generate maintenance instructions based on the correlation model between ash content and combustion chamber temperature gradient; B3. Dynamic environmental optimization: Automatically adjust the secondary air valve and urea injection amount according to the NOx concentration exceeding the standard.

7. The intelligent monitoring system for multi-source data fusion of a biomass smart power plant according to claim 6 is characterized by: The power regulation of the drying equipment in the intelligent feedback control module dynamically sets the target temperature range based on the real-time calculated compensation calorific value, and realizes closed-loop control in combination with the change of feed rate, wherein: The upper and lower limits of the target temperature range are positively correlated with the compensated calorific value; The feed rate is dynamically adjusted based on the deviation between the target temperature and the actual temperature. The greater the deviation, the greater the reduction in feed rate.

8. The intelligent monitoring system for multi-source data fusion of a biomass smart power plant according to claim 7 is characterized by: In the dual-channel deep residual network: Fuel characteristic channel: A one-dimensional convolution structure is used to extract the time series variation characteristics of calorific value and moisture content; Equipment operation channel: Multi-scale residual learning is performed on boiler temperature and flue gas oxygen content, and the output is fused and passed to the fully connected layer; The multi-scale residual learning includes parallel processing of device operation data by convolution kernels of at least three different scales.

9. The intelligent monitoring system for multi-source data fusion of a biomass smart power plant according to claim 8 is characterized by: The joint singular value decomposition in the coupling matrix reduction improves the sensitivity of combustion efficiency by introducing the equipment operation state weight matrix, where: The weight matrix is ​​dynamically calculated based on the product of boiler temperature and flue gas oxygen content; The total correlation coefficient between the number of principal components retained after order reduction and combustion efficiency is no less than 80%.

Images