Cogeneration heat supply cost calculation method and system based on BP neural network
Patent Information
- Application Number
- CN202610913193.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-06-24
- Publication Date
- 2026-09-15
Smart Images
Figure CN122760180A_ABST
Abstract
Description
Technical Field
[0001] The embodiments disclosed herein belong to the technical field of heating cost accounting for combined heat and power (CHP) units, specifically relating to a method and system for calculating CHP heating costs based on a BP neural network. Background Technology
[0002] Heating costs are a core indicator for the operation and management of combined heat and power (CHP) units and district heating companies, and their calculation results directly affect heating pricing, cost control, and corporate profit optimization. Heating costs are influenced by multiple factors, including heating load, equipment operating status, fuel prices, ambient temperature, and pipeline losses, exhibiting strong nonlinearity and dynamic changes. Traditional heating cost calculation methods have many technical shortcomings. Firstly, traditional methods often employ empirical formulas or itemized statistical methods. Empirical formulas are based on fixed operating conditions and cannot adapt to the full operating conditions of the unit under varying loads and environments, resulting in large calculation deviations. Itemized statistical methods only perform ex-post statistics on cost items such as fuel, water, electricity, and maintenance, which cannot achieve real-time dynamic calculations and do not consider the coupling effect of various influencing factors, resulting in low accounting accuracy.
[0003] Secondly, existing cost calculation methods do not quantify and screen influencing factors, and include irrelevant or weakly related factors in the calculation, which increases the computational complexity and reduces the accuracy of the calculation results. Although some methods introduce simple algorithms for cost fitting, the model structure is simple and cannot capture the nonlinear mapping relationship between multiple factors, resulting in poor adaptability.
[0004] Third, traditional methods are mostly offline post-event calculations, and data collection and processing are lagging behind. They cannot achieve real-time online calculation of heating costs, which makes it difficult to meet the needs of real-time scheduling of cogeneration units and dynamic cost control of heating companies. They also cannot provide accurate cost data support for the peak-shaving revenue accounting of cogeneration units under the thermal power spot market. Summary of the Invention
[0005] The embodiments disclosed herein aim to at least solve one of the technical problems existing in the prior art, and provide a method and system for calculating the heating cost of cogeneration based on BP neural network.
[0006] One aspect of this disclosure provides a method for calculating the heating cost of combined heat and power (CHP) systems based on a BP neural network, the method comprising: Real-time collection of multi-dimensional influencing factors of heating cost of cogeneration units, including unit operation indicators, external environment indicators, material consumption indicators and pipeline operation indicators, and normalization processing of the influencing factor indicators to form a standardized indicator set; The standardized index set is input into a pre-trained heating cost calculation model, and the real-time heating cost is calculated through forward propagation; wherein, the heating cost calculation model is obtained by training a BP neural network model based on historical operating data under all operating conditions; The real-time heating cost is sent to the unit's DCS system or the heating company's cost management platform. The heating cost calculation model is incrementally iteratively trained by collecting operational data at a preset period to update the model parameters.
[0007] Furthermore, the unit operating parameters include heating load, heating steam pressure, heating steam temperature, boiler heating efficiency, and turbine extraction steam rate; and / or, The external environmental indicators include outdoor ambient temperature, ambient humidity, and heating period; and / or, The material consumption indicators include the unit price of coal, the amount of coal consumed, the unit price of electricity, and the electricity consumption of auxiliary heating equipment; and / or, The pipeline operation indicators include the heating pipeline transmission distance, the pipeline heat loss rate, and the power consumption of the pipeline circulating water pump.
[0008] Furthermore, the real-time heating cost is the cost per unit heating load or the cost per unit time for heating.
[0009] Furthermore, the BP neural network model is pre-trained through the following steps: Collect historical operating data of the combined heat and power unit under all operating conditions, including historical data of influencing factors and actual calculation data of heating costs for the corresponding period. The historical running data is preprocessed and divided into training datasets and test datasets; The BP neural network model is iteratively trained using the training dataset until the loss function meets the preset error accuracy. The accuracy of the trained model is verified using the test dataset. The average relative error between the predicted heating cost and the actual heating cost is calculated. If the average relative error is greater than the preset relative error threshold, the model parameters are adjusted and the model is retrained until the accuracy requirements are met, thus obtaining the heating cost calculation model.
[0010] Furthermore, the BP neural network model includes an input layer, a hidden layer, and an output layer. The number of neurons in the input layer is consistent with the number of indicators in the standardized indicator set. The hidden layer uses the ReLU activation function, and the output layer uses a linear activation function.
[0011] Furthermore, the number of hidden layers is 1 or 2, and the number of neurons in the hidden layers is 1.5 to 2 times the number of neurons in the input layers.
[0012] Furthermore, the learning rate of the BP neural network model is 0.01~0.05, the number of iterations is 200~500, the error accuracy is ≤0.001, the loss function is mean squared error, and the optimization algorithm is gradient descent.
[0013] Another aspect of this disclosure provides a cogeneration heating cost calculation system based on a BP neural network, the system comprising: The indicator acquisition module is used to collect multi-dimensional influencing factors of heating cost of cogeneration units in real time, including unit operation indicators, external environment indicators, material consumption indicators and pipeline operation indicators. The influencing factor indicators are normalized to form a standardized indicator set. The cost calculation module is used to input the standardized index set into a pre-trained heating cost calculation model and calculate the real-time heating cost through forward propagation; wherein, the heating cost calculation model is obtained by training a BP neural network model based on historical operating data under all operating conditions; The result sending module is used to send the real-time heating cost to the unit's DCS system or the heating company's cost management platform. The incremental update module is used to collect operating data at a preset period to perform incremental iterative training on the heating cost calculation model and update the model parameters.
[0014] Another aspect of this disclosure provides an electronic device, comprising: At least one processor; and, A memory communicatively connected to the at least one processor is used to store one or more programs, which, when executed by the at least one processor, enable the at least one processor to implement the above-described method for calculating the cost of cogeneration heating based on a BP neural network.
[0015] Another aspect of this disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for calculating the cost of cogeneration heating based on a BP neural network.
[0016] This disclosure discloses a method and system for calculating heating costs in combined heat and power (CHP) systems based on BP neural networks. Through screening of multi-dimensional strongly correlated influencing factors and nonlinear fitting of the BP neural network, the model calculates heating costs with a small average relative error and accurate cost accounting results. It achieves real-time online acquisition, calculation, and output of heating costs with a short response time, overcoming the lag of traditional offline post-calculation. This provides accurate and real-time cost data support for real-time scheduling of CHP units and peak-shaving revenue calculation in the thermal power spot market. An incremental iterative optimization mechanism is established, which can continuously optimize based on the dynamic operating data of the units. The updated model adapts to factors such as equipment aging, changes in pipeline losses, and fuel price fluctuations, ensuring stable long-term calculation accuracy without frequent manual parameter adjustments. No new hardware is required; software modules can be developed directly on existing DCS systems of cogeneration units or cost management platforms of heating companies, resulting in low retrofit costs, simple operation, and applicability to various condensing, back-pressure, and extraction-condensing cogeneration units and district heating companies, making it highly scalable. It provides comprehensive and accurate data support for cost analysis, cost control strategy formulation, and heating pricing optimization for heating companies, improving their operational efficiency. Attached Figure Description
[0017] Figure 1 This is a flowchart illustrating a method for calculating the heating cost of cogeneration based on a BP neural network according to an embodiment of this disclosure. Figure 2 This is a schematic diagram of the structure of a cogeneration heating cost calculation system based on a BP neural network according to another embodiment of the present disclosure; Figure 3 This is a schematic diagram of the structure of an electronic device according to another embodiment of the present disclosure. Detailed Implementation
[0018] The technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this disclosure, and not all of them. Based on the embodiments of this disclosure, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this disclosure.
[0019] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this disclosure.
[0020] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.
[0021] It should be understood that although the terms first, second, third, etc., may be used in this disclosure to describe various components, these components should not be limited by these terms. These terms are used to distinguish one component from another. Therefore, the first component discussed below may be referred to as the second component without departing from the teachings of this disclosure. As used in this disclosure, the term "and / or" includes all combinations of any one and more of the associated listed items.
[0022] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of exemplary embodiments, and the modules or processes in the drawings are not necessarily necessary for implementing this disclosure, and therefore cannot be used to limit the scope of protection of this disclosure.
[0023] like Figure 1 As shown, one embodiment of this disclosure provides a method for calculating the heating cost of combined heat and power (CHP) based on a BP neural network, including: Step S1: Collect multi-dimensional influencing factors of heating cost of cogeneration unit in real time, including unit operation indicators, external environment indicators, material consumption indicators and pipeline operation indicators, and normalize the influencing factor indicators to form a standardized indicator set.
[0024] Specifically, a multi-dimensional indicator system for influencing heating costs was constructed, and heating cost accounting items were determined. Focusing on the entire heating process of combined heat and power (CHP) units, core influencing factors strongly correlated with heating costs were identified through correlation analysis and categorized into four main types: unit operation, external environment, material consumption, and pipeline operation. A real-time, quantifiable indicator system was constructed. Simultaneously, core heating cost accounting items were clearly defined to ensure the completeness and standardization of cost calculations.
[0025] Key influencing factors include: Unit operating parameters: real-time heating load (t / h), heating steam pressure (MPa), heating steam temperature (°C), boiler heating efficiency (%), turbine extraction steam rate (t / h); External environmental indicators: outdoor ambient temperature (°C), ambient humidity (%), heating period (peak / average / valley); Material consumption indicators: coal price (yuan / ton), coal consumption (ton / h), electricity price (yuan / kWh), and power consumption of auxiliary heating equipment (kWh / h). Pipeline operation indicators: heating pipeline transmission distance (km), pipeline heat loss rate (%), pipeline circulating water pump power consumption (kWh / h).
[0026] Core heating cost accounting items include: The costs include coal-fired heating costs, electricity heating costs, equipment maintenance and depreciation costs, pipeline operation and maintenance costs, and labor and other fixed costs; among which, labor and other fixed costs are allocated to the unit time cost based on the heating duration.
[0027] All the aforementioned influencing factors were collected in real time and normalized to eliminate the impact of different dimensions and orders of magnitude on the model, forming a standardized indicator set. Cost accounting items were uniformly quantified and converted into unit heating load cost (yuan / GJ) or unit time heating cost (yuan / h), which were used as model output labels.
[0028] Step S2: Input the standardized index set into the pre-trained heating cost calculation model, and calculate the real-time heating cost through forward propagation.
[0029] Specifically, the heating cost calculation model is obtained by training a BP neural network model based on historical operating data under all operating conditions. According to the number and nonlinear characteristics of the factors influencing heating costs, a three-layer BP neural network model (input layer - hidden layer - output layer) is constructed. The model structure adapts to the mapping relationship between heating costs and multi-dimensional influencing factors. The specific parameters are designed as follows: 1. Input layer: The number of neurons is the same as the number of indicators in the standardized indicator set in step S1. Each neuron corresponds to a core influencing factor indicator and is used to receive standardized real-time indicator data. 2. Hidden Layers: Set 1 to 2 hidden layers, with the number of neurons being 1.5 to 2 times that of the input layer neurons. Use the ReLU activation function to extract and map multi-factor nonlinear features and avoid model overfitting. 3. Output layer: The number of neurons is 1. The output is the heating cost calculation result, that is, the unit heating load cost (yuan / GJ) or the unit time heating cost (yuan / h). A linear activation function is used to ensure the quantitative accuracy of the output result. 4. Model parameters: The learning rate is set to 0.01~0.05, the number of iterations is set to 200~500, the error accuracy threshold is set to ≤0.001, the gradient descent method is used for model optimization, and the mean squared error (MSE) is used as the model loss function to measure the deviation between the predicted value and the actual value.
[0030] Train the constructed BP neural network model: 1. Collect historical operating data of the cogeneration unit under all operating conditions for the past 6 to 12 months, including historical data of all influencing factor indicators, as well as actual accounting data of heating costs for the corresponding period. Ensure that the dataset covers different heating conditions such as full load, partial load, and low load of the unit, as well as operating scenarios with different ambient temperatures and different fuel prices. The dataset sample size is no less than 5,000 sets. 2. Preprocess the collected raw data (i.e., historical operating data under all operating conditions) to remove outliers and missing values, and interpolate and complete any missed data to ensure the integrity and validity of the data; 3. The standardized index sample set (input features) obtained after preprocessing and the actual heating cost accounting data (output labels) are divided into training dataset and test dataset in a 7:3 ratio. The training dataset is used for iterative training of the model, and the test dataset is used for accuracy verification of the model. 4. Input the training dataset into the constructed BP neural network model, and iterate the model training according to the preset learning rate and number of iterations. Calculate the predicted heating cost through forward propagation, and correct the model weights and biases through backpropagation until the model loss function value is no greater than the error accuracy threshold (0.001). 5. Input the test dataset into the model trained in the previous step to verify the model's accuracy. Calculate the average relative error between the predicted and actual heating costs output by the model. If the average relative error is ≤5%, the model is deemed to meet the accuracy requirements. If the average relative error is >5%, adjust the number of neurons in the hidden layer, the learning rate, and other parameters of the model, and retrain the model until the accuracy requirements are met, thus obtaining the heating cost calculation model.
[0031] The standardized index set from step S1 is input into the trained heating cost calculation model. The model quickly outputs the real-time heating cost calculation results through forward propagation, with a calculation response time of ≤10s, thus realizing real-time online accounting of heating costs.
[0032] Step S3: Send the real-time heating cost to the unit's DCS system or the heating company's cost management platform.
[0033] Specifically, the real-time heating cost data (including the composition of cost items, the contribution of influencing factors, etc.) obtained in step S2 is output to the unit's DCS system or the heating company's cost management platform in real time, and historical cost data is automatically stored to provide data support for cost analysis and control strategy formulation.
[0034] Step S4: Collect operating data at a preset cycle to perform incremental iterative training on the heating cost calculation model and update the model parameters.
[0035] Specifically, on a monthly basis, the unit's monthly operating data and actual cost accounting data are collected and added to the model training dataset. The BP neural network model is then incrementally iteratively trained, and the model weights and biases are updated to adapt the model to dynamic factors such as unit equipment aging, fuel price fluctuations, and changes in pipeline operating conditions, ensuring the long-term stability of the model's calculation accuracy.
[0036] The following describes the specific implementation of this disclosure in detail, taking a 600MW supercritical extraction-condensing cogeneration unit as an example: I. Basic Parameters 1. Unit parameters: 600MW supercritical extraction condensing cogeneration unit, heating load range 50t / h~250t / h, heating steam pressure 1.6MPa, heating steam temperature 350℃, boiler heating efficiency 92%~95%, heating pipeline transmission distance 15km; 2. BP neural network model parameters: 12 neurons in the input layer (corresponding to 12 core influencing factors), 1 hidden layer with 20 neurons, and 1 neuron in the output layer; learning rate 0.03, number of iterations 300, error accuracy threshold 0.001, activation function is ReLU (hidden layer) and linear function (output layer), and loss function is mean squared error (MSE). 3. Cost accounting objective: Calculate the total cost of heating per unit time (yuan / h). The core accounting items include coal cost, electricity cost, equipment maintenance and depreciation cost, pipeline operation and maintenance cost, labor and other fixed costs (averaged at 200 yuan / h per hour).
[0037] II. Implementation Preparation 1. Data acquisition module deployment: Based on the existing DCS system of the unit, a data acquisition module is developed to collect data on 12 core influencing factors in real time at a frequency of 5 seconds / time. At the same time, it is connected to the coal and electricity price management system to realize real-time synchronization of material price data. 2. Dataset Construction: Collect historical operating data of the unit under all operating conditions for nearly 8 months, totaling 8,000 valid samples. After removing outliers and missing values, the index data were normalized by min-max, and the actual heating cost data was quantitatively calculated and converted into the total heating cost per unit time (yuan / h). The data was divided into 5,600 training samples and 2,400 test samples in a 7:3 ratio. 3. BP Neural Network Model Construction: Based on the Python / TensorFlow framework, a three-layer BP neural network model was built on the unit cost management platform, and the model was initialized according to the preset parameters; 4. Model Training and Validation: The training samples were input into the model for iterative training. After 300 iterations, the model loss function value dropped to 0.0008, which met the error accuracy requirements. The test samples were input into the model, and the average relative error of the model calculation was verified to be 3.8% < 5%, which met the requirements of practical applications.
[0038] III. Implementation Steps 1. Real-time data acquisition and preprocessing: The data acquisition module collects data on 12 core influencing factors such as unit heating load, ambient temperature, coal consumption, and pipeline loss rate at a frequency of 5 seconds / time. It automatically removes outliers and normalizes the data to form a standardized real-time indicator set. 2. Real-time heating cost calculation: The standardized real-time index set is input into the trained BP neural network model. The model quickly calculates the total heating cost per unit time through forward propagation and outputs the result within 8 seconds. For example, when the heating load is 200t / h, the outdoor ambient temperature is -5℃, and the coal price is 1000 yuan / h, the model calculates the total real-time heating cost to be 18650 yuan / h, including 15200 yuan / h for coal, 1800 yuan / h for electricity, 850 yuan / h for equipment maintenance and depreciation, 600 yuan / h for pipeline operation and maintenance, and 200 yuan / h for labor and other fixed costs. 3. Output and storage of calculation results: The total real-time heating cost and the breakdown of cost items are output to the unit's DCS system and the heating company's cost management platform in real time. At the same time, the data and the corresponding influencing factor index data are automatically stored to form a historical cost database. 4. Model Iterative Optimization: At the end of each month, real-time operation data and actual cost accounting data of approximately 2,000 units are collected and added to the model training dataset. The BP neural network model is then incrementally iteratively trained, and the model weights and biases are updated to ensure that the model adapts to factors such as fuel price fluctuations and changes in equipment operating status. After training, the model's calculation accuracy is maintained at an average relative error of ≤4%. 5. Cost Analysis and Application: Heating companies can conduct cost analysis based on the real-time cost data, cost item composition, and historical database output by the model, identify core cost control points such as coal cost and pipeline losses, and formulate targeted cost control strategies. At the same time, the real-time cost data provides accurate support for the calculation of peak-shaving revenue of thermal power units in the spot market, realizing dynamic quantitative analysis of "revenue-cost".
[0039] IV. Implementation Results In this embodiment, the cogeneration heating cost calculation method based on BP neural network disclosed herein is adopted, which significantly improves the unit heating cost accounting effect: 1. Significantly improved calculation accuracy: The average relative error of heating cost calculation is stable at 3.5%~4.0%, which is more than 65% lower than the traditional empirical formula method (average relative error 10%~15%), and the calculation results are highly consistent with the actual values; 2. Real-time online accounting: The model calculation response time is ≤8s, which can track the dynamic changes of factors such as heating load and fuel price in real time and output real-time cost data, completely solving the problem of lag in traditional offline post-event accounting; 3. Adaptable to all operating conditions: Under all operating conditions and multiple scenarios with unit heating load of 50t / h to 250t / h and outdoor ambient temperature of -15℃ to 10℃, the model can accurately calculate heating costs without significant deviation and has strong adaptability. 4. Facilitating Cost Control and Revenue Optimization: By providing real-time output of cost component data, core cost control points can be accurately identified, reducing the overall heating cost of the unit by 5% to 8%. Simultaneously, real-time cost data provides precise support for the formulation of peak-shaving strategies in the thermal power spot market, increasing the overall peak-shaving revenue of the unit by 3% to 5%. 5. Low deployment cost and easy operation: Software modules can be developed directly on the existing DCS system and cost management platform without the need for new hardware, resulting in low transformation costs. Model iteration and optimization are completed automatically without frequent manual intervention, thus reducing operation and management costs.
[0040] This disclosure discloses a method for calculating the heating cost of cogeneration based on a BP neural network. Through screening of multi-dimensional strongly correlated influencing factors and nonlinear fitting of the BP neural network, the average relative error between the model's calculated heating cost and the actual calculated value is small, resulting in more accurate cost accounting. The calculation model is trained based on historical data of the unit under all operating conditions, adapting to dynamic changes in heating load, ambient temperature, fuel prices, and other factors. It covers all heating conditions, including full load, partial load, and low load, solving the problem that traditional methods are only applicable to fixed operating conditions. It achieves real-time online acquisition, calculation, and output of heating costs with a short response time, overcoming the lag of traditional offline post-calculation. This provides accurate and real-time cost data for real-time scheduling of cogeneration units and peak-shaving revenue calculation under the thermal power spot market. Supported by an incremental iterative optimization mechanism, the model is continuously updated based on dynamic operating data of the unit, adapting to factors such as equipment aging, changes in pipeline losses, and fuel price fluctuations. This ensures the long-term stability of the model's calculation accuracy without the need for frequent manual parameter adjustments. No new hardware is required; software modules can be developed directly on the existing DCS system of the cogeneration unit or the cost management platform of the heating company. The transformation cost is low, the operation is simple, and it is applicable to various condensing, back-pressure, and extraction-condensing cogeneration units and various district heating companies, making it highly scalable. It can output data such as cost item composition and the contribution of influencing factors in real time, and automatically store historical cost data, providing comprehensive and accurate data support for cost analysis, cost control strategy formulation, and heating pricing optimization for heating companies, thereby improving the efficiency of enterprise operation and management.
[0041] like Figure 2 As shown, another embodiment of this disclosure provides a cogeneration heating cost calculation system based on a BP neural network, including: The indicator acquisition module 210 is used to collect multi-dimensional influencing factors of heating cost of cogeneration unit in real time, including unit operation indicators, external environment indicators, material consumption indicators and pipeline operation indicators, and to normalize the influencing factor indicators to form a standardized indicator set. The cost calculation module 220 is used to input the standardized index set into a pre-trained heating cost calculation model and calculate the real-time heating cost through forward propagation; wherein, the heating cost calculation model is obtained by training a BP neural network model based on historical operating data under all operating conditions; The result sending module 230 is used to send the real-time heating cost to the unit's DCS system or the heating company's cost management platform. The incremental update module 240 is used to collect operating data at a preset period to perform incremental iterative training on the heating cost calculation model and update the model parameters.
[0042] Specifically, the cogeneration heating cost calculation system based on BP neural network of this disclosure is used to implement the cogeneration heating cost calculation method based on BP neural network described in the above embodiments. The specific implementation process has been described in detail in the above embodiments and will not be repeated here.
[0043] like Figure 3 As shown, another embodiment of this disclosure provides an electronic device, including: At least one processor 301; and a memory 302 communicatively connected to the at least one processor 301 for storing one or more programs that, when executed by the at least one processor 301, enable the at least one processor 301 to implement the above-described method for calculating the cost of cogeneration heating based on a BP neural network.
[0044] The memory 302 and processor 301 are connected via a bus, which can include any number of interconnecting buses and bridges. The bus connects various circuits of one or more processors 301 and memory 302 together. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. A bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 301 is transmitted over a wireless medium via an antenna, which further receives data and transmits it to processor 301.
[0045] Processor 301 is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory 302 can be used to store data used by processor 301 during operation.
[0046] Another embodiment of this disclosure provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for calculating the heating cost of cogeneration based on a BP neural network.
[0047] The computer-readable storage medium may be included in the systems or electronic devices disclosed herein, or it may exist independently.
[0048] Computer-readable storage media can be any tangible medium that contains or stores a program, and can be an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device. More specific examples include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, optical fibers, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.
[0049] Computer-readable storage media may also include data signals propagated in baseband or as part of a carrier wave, carrying computer-readable program code, specific examples of which include, but are not limited to, electromagnetic signals, optical signals, or any suitable combination thereof.
[0050] It is understood that the above embodiments are merely exemplary embodiments used to illustrate the principles of this disclosure, and this disclosure is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and substance of this disclosure, and these modifications and improvements are also considered to be within the scope of protection of this disclosure.
Claims
1. A method for calculating the heating cost of combined heat and power (CHP) based on a BP neural network, characterized in that, The method includes: Real-time collection of multi-dimensional influencing factors of heating cost of cogeneration units, including unit operation indicators, external environment indicators, material consumption indicators and pipeline operation indicators, and normalization processing of the influencing factor indicators to form a standardized indicator set; The standardized index set is input into a pre-trained heating cost calculation model, and the real-time heating cost is calculated through forward propagation; wherein, the heating cost calculation model is obtained by training a BP neural network model based on historical operating data under all operating conditions; The real-time heating cost is sent to the unit's DCS system or the heating company's cost management platform. The heating cost calculation model is incrementally iteratively trained by collecting operational data at a preset period to update the model parameters.
2. The method according to claim 1, characterized in that, The unit operating parameters include heating load, heating steam pressure, heating steam temperature, boiler heating efficiency, and turbine extraction steam rate; and / or, The external environmental indicators include outdoor ambient temperature, ambient humidity, and heating period; and / or, The material consumption indicators include the unit price of coal, the amount of coal consumed, the unit price of electricity, and the electricity consumption of auxiliary heating equipment; and / or, The pipeline operation indicators include the heating pipeline transmission distance, the pipeline heat loss rate, and the power consumption of the pipeline circulating water pump.
3. The method according to claim 1, characterized in that, The real-time heating cost is the cost per unit heating load or the cost per unit time for heating.
4. The method according to any one of claims 1 to 3, characterized in that, The BP neural network model is pre-trained through the following steps: Collect historical operating data of the combined heat and power unit under all operating conditions, including historical data of influencing factors and actual calculation data of heating costs for the corresponding period. The historical running data is preprocessed and divided into training datasets and test datasets; The BP neural network model is iteratively trained using the training dataset until the loss function meets the preset error accuracy. The accuracy of the trained model is verified using the test dataset. The average relative error between the predicted heating cost and the actual heating cost is calculated. If the average relative error is greater than the preset relative error threshold, the model parameters are adjusted and the model is retrained until the accuracy requirements are met, thus obtaining the heating cost calculation model.
5. The method according to claim 4, characterized in that, The BP neural network model includes an input layer, a hidden layer, and an output layer. The number of neurons in the input layer is the same as the number of indicators in the standardized indicator set. The hidden layer uses the ReLU activation function, and the output layer uses a linear activation function.
6. The method according to claim 5, characterized in that, The number of hidden layers is 1 or 2, and the number of neurons in the hidden layers is 1.5 to 2 times the number of neurons in the input layers.
7. The method according to claim 4, characterized in that, The learning rate of the BP neural network model is 0.01~0.05, the number of iterations is 200~500, the error accuracy is ≤0.001, the loss function is mean squared error, and the optimization algorithm is gradient descent.
8. A cogeneration heating cost calculation system based on BP neural network, characterized in that, The system includes: The indicator acquisition module is used to collect multi-dimensional influencing factors of heating cost of cogeneration units in real time, including unit operation indicators, external environment indicators, material consumption indicators and pipeline operation indicators. The influencing factor indicators are normalized to form a standardized indicator set. The cost calculation module is used to input the standardized index set into a pre-trained heating cost calculation model and calculate the real-time heating cost through forward propagation; wherein, the heating cost calculation model is obtained by training a BP neural network model based on historical operating data under all operating conditions; The result sending module is used to send the real-time heating cost to the unit's DCS system or the heating company's cost management platform. The incremental update module is used to collect operating data at a preset period to perform incremental iterative training on the heating cost calculation model and update the model parameters.
9. An electronic device, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor is used to store one or more programs, which, when executed by the at least one processor, enable the at least one processor to implement the cogeneration heating cost calculation method based on a BP neural network as described in any one of claims 1 to 7.
10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the method for calculating the heating cost of cogeneration based on a BP neural network as described in any one of claims 1 to 7.