A holographic see-through system and method for monitoring combustion in a coal-fired boiler furnace

Abstract

The application discloses a kind of holographic perspective systems and methods for monitoring combustion of coal-fired boiler furnace, comprising: actual coal-fired boiler and furnace holographic perspective platform, the model of this holographic perspective platform includes: using combustion model to calculate the flow field, temperature field and composition field of each place in the combustion process of actual coal-fired boiler furnace and carry out demonstration combustion process monitoring module, and the output of combustion model is calculated with the operation instruction of operating personnel and is demonstrated combustion process prediction module.For combustion model, it includes temperature field model, flow field model and composition field model of coal combustion in furnace based on computational fluid dynamics and deep learning method.Furnace holographic perspective platform obtains operating data from thermal power generating unit for processing, calculates the combustion state of furnace using combustion model, sends the calculation result to man-machine interaction unit for demonstration, and can also predict the combustion process of furnace according to the control command of man-machine interaction unit, and demonstrate the prediction result on man-machine interface.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of operation process monitoring of thermal power generating units, and particularly relates to a holographic perspective system for monitoring combustion in a furnace of a coal-fired boiler. BACKGROUND

[0002] Adjustment of combustion in the furnace is the core of operation of the boiler, and is directly related to safety, economy and environmental protection of the entire unit. The thermal power unit needs to be frequently adjusted in actual operation process, such as rise and fall of unit load, and adjustment of the coal pulverizing system. Changes of these working conditions will directly affect changes of the combustion state in the furnace of the boiler. Due to the complexity of the coal combustion process in the furnace, there has been a lack of scientific criteria for combustion adjustment. The existing adjustment mainly relies on experience and habits of the operating personnel, and usually lacks scientific basis. More scientific adjustment of the combustion in the furnace of the boiler is helpful to improve the operation level of the thermal power unit and improve safety and economy of operation of the unit.

[0003] Some methods for monitoring the flame in the furnace are also disclosed in the prior art. Chinese patent CN216976899U discloses a holographic perspective flame imaging device, in which an optical imaging device is used to image the flame of a fireplace. Chinese patent CN115758875A discloses a digital twin method for a boiler combustion chamber, which combines a computational fluid dynamics method and a neural network method to calculate the temperature, pressure, velocity and composition of the boiler combustion chamber.

[0004] However, the above prior art cannot guarantee that the calculation results of the furnace related field are consistent with the actual situation on site, and there is a large deviation between the calculation results of the model and the actual situation, and the design is not updated when the deviation is large, which cannot well guide the combustion adjustment on site.

[0005] Invention objectives

[0006] The present application aims to solve the problems in the prior art, and provides a holographic perspective system and method for monitoring combustion in a furnace of a coal-fired boiler, which can directly monitor the combustion process of the thermal power unit and scientifically predict possible consequences of the combustion process. SUMMARY

[0007] According to one aspect of the present application, a holographic perspective system for monitoring combustion in a furnace of a coal-fired boiler is provided, which is composed of an actual coal-fired unit and a holographic perspective platform.

[0008] The holographic perspective platform obtains the current operation data and operation instructions of the actual coal-fired unit from the distributed control system (DCS) of the actual coal-fired unit, including operation parameters and furnace measurement parameters; if the furnace cross-section temperature measurement equipment output of the actual coal-fired unit is not connected to the DCS, the temperature measurement results are obtained directly from the interface provided by the temperature measurement equipment, and at this time the temperature measurement results of the actual coal-fired boiler furnace observation hole are manually entered into the holographic perspective platform by artificial means;

[0009] The functions of the furnace holographic perspective platform include modeling, monitoring and prediction mode switching and model updating; the modeling refers to modeling the furnace combustion process of the actual coal-fired unit, and the established model is composed of a computational fluid dynamics model and a deep learning model; the operation modes of the furnace holographic perspective platform include a monitoring mode and a prediction mode, wherein the monitoring mode displays the current furnace combustion process in real time according to the actual boiler operation parameters; the prediction mode displays the combustion process of the boiler when running at the parameter set by the human being, and the two modes can be switched by artificial means.

[0010] Preferably, the computational fluid dynamics model adopts a partition model for calculation, specifically, the boiler furnace is divided into different regions according to the furnace cross-section temperature measurement positions, and the computational grids of the computational fluid dynamics models of the furnace temperature field, flow field and composition field are established respectively; first, the calculation starts from the region where the burner is located, and the boundary conditions are based on the actual boiler operation parameters; then the adjacent region of the combustion region is calculated, and the calculation results of the combustion region are used as the boundary conditions; and the other furnace regions are calculated in the same way.

[0011] Further preferably, the computational fluid dynamics model is used to calculate the temperature field, flow field and composition field in the furnace of the actual coal-fired unit under different coal mill combination operation modes and typical load conditions; based on the calculation results, the deep learning-based combustion process models of the furnace temperature field, flow field and composition field are established respectively; wherein the deep learning model can adopt a three-dimensional convolution network structure, the input of the combustion process model is the operation parameters of the boiler, and the output is the numerical values of the furnace temperature field, flow field and composition field.

[0012] Preferably, when the furnace holographic perspective platform is in the monitoring mode, the furnace holographic perspective platform model takes the current operation data and operation instructions as input, calculates the temperature field, flow field and composition field of the furnace, and then displays the calculation results in the form of images, specifically, the results are displayed in two dimensions in the form of the cross section, vertical section in front and back, and vertical section in left and right directions at the positions specified by the operation personnel, or the above-mentioned sections are combined to realize three-dimensional display; when the platform is in the prediction mode, the model takes the set operation parameters manually input by the operation personnel in the platform as input, calculates the corresponding furnace temperature field, flow field and composition field, and displays them in the above-mentioned two-dimensional or three-dimensional way.

[0013] Preferably, the model update indicates that the furnace holographic perspective platform has a model automatic update function; when the furnace holographic perspective platform is in a monitoring mode, if the deviation of the current model calculation result from the actual cross-section temperature measurement and the tube fire hole temperature measurement result exceeds the set threshold value, the combustion process model is automatically updated, and specifically, the average variance of the actual cross-section temperature measurement, the tube fire hole temperature measurement result and the combustion process model calculation value or other type loss function is minimized as the training target, and the combustion process model is fine-tuned.

[0014] According to another aspect of the present application, a method for monitoring the combustion of a coal-fired boiler furnace by using the above-mentioned holographic perspective system is provided, characterized in that it comprises the following steps:

[0015] Step 1: obtaining the current operation data and operation instructions of the actual coal-fired unit from the DCS of the actual coal-fired unit by the holographic perspective platform, including operation parameters and furnace measurement parameters; if the furnace cross-section temperature measurement equipment output of the actual coal-fired unit is not connected to the DCS, the temperature measurement result is obtained directly from the interface provided by the temperature measurement equipment, and at this time the temperature measurement result of the actual coal-fired boiler furnace observation hole is manually entered into the holographic perspective platform by artificial means;

[0016] Step 2: modeling the furnace combustion process of the actual coal-fired unit, the model is composed of a computational fluid dynamics model and a deep learning model; the computational fluid dynamics model uses a partition model for calculation, specifically the furnace cross-section temperature measurement position is divided into different regions, and the computational grid of the computational fluid dynamics model of the furnace temperature field, flow field and composition field is established respectively; first, the calculation starts from the area where the burner is located, and the boundary condition is based on the actual boiler operation parameters; then the adjacent area of the combustion area is calculated, and the calculation result of the combustion area is used as the boundary condition; other furnace areas are calculated in the same way;

[0017] Step 3: using the computational fluid dynamics model to calculate the temperature field, flow field and composition field in the furnace of the actual coal-fired unit under different coal mill combination operation modes and typical load conditions; based on the calculation results, the deep learning-based combustion process model of the furnace temperature field, flow field and composition field is established respectively; the deep learning model can adopt a three-dimensional convolution network structure, the input of the combustion process model is the operation parameters of the boiler, and the output is the numerical value of the furnace temperature field, flow field and composition field;

[0018] Step 4, when the furnace holographic perspective platform is in the monitoring mode, if the deviation of the current model calculation result and the actual cross-section temperature measurement and the temperature measurement result of the tube fire hole exceeds the set threshold value, the combustion process model is automatically updated, and the average variance or other type loss function of the actual cross-section temperature measurement, the temperature measurement result of the tube fire hole and the combustion process model calculation value is minimized as the training target, and the combustion process model is fine-tuned. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 is the architecture schematic diagram of the coal-fired boiler furnace combustion holographic perspective system of the present application.

[0020] Figure 2 is the three-dimensional structure display diagram of the coal-fired boiler furnace combustion holographic perspective system of the present application. DETAILED DESCRIPTION

[0021] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0022] Those skilled in the art should understand that the step numbers used herein are only for the convenience of description, and are not limited to the execution sequence of the steps. The terms used in the present application are only for the purpose of describing the specific embodiments and are not intended to limit the present application. As used in the present application and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" can include the plural forms. The term "and / or" refers to any combination of one or more of the associated listed items and all possible combinations, and includes these combinations. In the description of the present application, the meaning of "a plurality of" is at least two, for example, two, three, etc., unless otherwise specifically limited.

[0023] Figure 1 is the architecture schematic diagram of the coal-fired boiler furnace combustion holographic perspective system of the present application. As shown in the figure, the coal-fired boiler furnace combustion holographic perspective system is composed of an actual coal-fired unit and a furnace holographic perspective platform. The furnace holographic perspective platform obtains the current operation data and operation instructions of the unit from the DCS of the actual coal-fired unit. If the cross-section temperature measurement equipment output is not connected to the DCS, the temperature measurement result is obtained directly from the interface provided by the temperature measurement equipment. The temperature measurement result of the observation fire hole is manually entered by artificial means.

[0024] The furnace holographic perspective platform includes modeling, monitoring and prediction mode modes and model updating functions.

[0025] The model of the furnace holographic perspective platform is composed of a computational fluid dynamics model and a deep learning model. The computational fluid dynamics model is calculated by using a partition model. Specifically, the furnace is divided into different regions according to the temperature measurement positions of the furnace cross section, and the calculation grids of the computational fluid dynamics models of the temperature field, flow field and composition field of the furnace are established respectively. First, the calculation starts from the region where the burner is located, and the boundary conditions are based on the actual operation parameters of the boiler; then the adjacent region of the combustion region is calculated, and the calculation results of the combustion region are used as the boundary conditions; and the other furnace regions are calculated in the same way. The temperature field, flow field and composition field in the furnace under different coal mill combination operation modes and typical load conditions are calculated according to the above method. Based on the calculation results, the deep learning-based combustion process models of the temperature field, flow field and composition field of the furnace are established respectively; the deep learning model can adopt a three-dimensional convolutional network structure, the input of the combustion process model is the operation parameters of the boiler, and the output is the numerical values of the temperature field, flow field and composition field of the furnace.

[0026] The mode of the furnace holographic perspective platform can be switched between monitoring and prediction modes as needed. When the platform is in the monitoring mode, the model takes the current operation data and operation instructions as the input, calculates the temperature field, flow field and composition field of the furnace, and then displays the calculation results in the form of images. Specifically, the results are displayed in two dimensions in the form of the cross section, the vertical section in the front-back direction and the vertical section in the left-right direction specified by the operator, or the above-mentioned sections are combined to realize three-dimensional display. Figure 2 is a three-dimensional structure display diagram of the coal-fired boiler furnace combustion holographic perspective system according to the present application, as shown in Figure 2 , wherein 1 is a cross section, 2 is a vertical section in the front-back direction, and 3 is a vertical section in the left-right direction. When the platform is in the prediction mode, the model takes the set operation parameters manually input by the operator in the platform as the input, calculates the corresponding temperature field, flow field and composition field of the furnace, and displays them in the above-mentioned two-dimensional or three-dimensional manner.

[0027] The furnace holographic perspective platform has an automatic model updating function. When the platform is in the monitoring mode, if the deviation of the current model calculation result from the actual cross section temperature measurement and the temperature measurement result of the tube fire hole exceeds the set threshold, the combustion process model is automatically updated. The update of the model is specifically to take the average variance of the actual temperature measurement result and the calculation value of the loss function model or the minimum of other types of loss function as the training target, and fine-tune the combustion process model. In order to prevent the combustion process model from deviating greatly from the actual process under other operating conditions due to fine-tuning, the early stopping and other deep learning methods are used to reduce the adverse effects of fine-tuning.

[0028] The embodiment also provides a method for monitoring the combustion of a coal-fired boiler furnace by using the above-mentioned holographic perspective system, which comprises the following steps:

[0029] Step 1, obtaining the current operation data and operation instructions of the actual coal-fired unit from the DCS of the actual coal-fired unit through the holographic perspective platform, including operation operation parameters and furnace measurement parameters; if the furnace cross-section temperature measurement equipment output of the actual coal-fired unit is not connected to the DCS, the temperature measurement results are obtained directly from the interface provided by the temperature measurement equipment, at this time the temperature measurement results of the actual coal-fired boiler furnace observation hole are manually entered into the holographic perspective platform by artificial means;

[0030] Step 2, modeling the furnace combustion process of the actual coal-fired unit, the model is composed of a computational fluid dynamics model and a deep learning model; the computational fluid dynamics model adopts a partition model for calculation, specifically the boiler furnace is divided into different regions according to the furnace cross-section temperature measurement position, and the calculation grid of the computational fluid dynamics model of the furnace temperature field, flow field and composition field is established respectively; first, the calculation starts from the area where the burner is located, and the boundary condition is based on the actual boiler operation parameters; then the adjacent area of the combustion area is calculated, and the calculation result of the combustion area is taken as the boundary condition; other furnace areas are calculated in the same way;

[0031] Step 3, using the computational fluid dynamics model to calculate the temperature field, flow field and composition field in the furnace of the actual coal-fired unit under different coal mill combination operation modes and typical load conditions; based on the calculation results, the deep learning-based combustion process model of the furnace temperature field, flow field and composition field is established respectively; the deep learning model can adopt a three-dimensional convolution network structure, the input of the combustion process model is the operation parameters of the boiler, and the output is the numerical value of the furnace temperature field, flow field and composition field;

[0032] Step 4, when the furnace holographic perspective platform is in the monitoring mode, if the deviation between the current model calculation result and the actual cross-section temperature and the temperature measurement result of the tube observation hole exceeds the set threshold, the combustion process model is automatically updated, specifically the average variance or other type loss function of the actual cross-section temperature, the temperature measurement result of the tube observation hole and the combustion process model calculation value is taken as the training target, and the combustion process model is fine-tuned.

[0033] Compared with the prior art, the present application has the following beneficial effects:

[0034] The present application can provide technical support for the actual operation of a thermal power unit boiler, and specifically reflects that when combustion adjustment, boiler load increase and decrease, and reverse grinding processes are performed, the operator can know in advance the influence of the relevant operation on the furnace combustion. In the prior art, when the thermal power unit is in variable load operation or reverse grinding due to the need for maintenance of the pulverizing system, the operator usually needs to adjust the furnace combustion. Such adjustment usually depends on the experience and habits of the operator, and there is room for optimization. The present application proposes a computational fluid dynamics model partition modeling method and calculation steps based on cross-section temperature measurement results, and proposes a deep learning network to ensure that the calculation results of the computational fluid dynamics model are consistent with the actual furnace combustion state, and also provides a specific scheme for the operator of the thermal power plant to use the platform. Therefore, the present application can improve the operation level of the operator and reduce the adverse effects of improper adjustment on the boiler.

Claims

1. A holographic see-through system for monitoring combustion in a furnace of a coal-fired boiler, characterized by, The holographic perspective system is composed of an actual coal-fired unit and a holographic perspective platform; The holographic perspective platform obtains current operation data and operation instructions of the actual coal-fired unit from a distributed control system (DCS) of the actual coal-fired unit, including operation operation parameters and furnace measurement parameters; if the output of the furnace cross-section temperature measurement equipment of the actual coal-fired unit is not connected to the DCS, the temperature measurement result is obtained from the interface provided by the temperature measurement equipment; at this time, the temperature measurement result of the actual coal-fired boiler furnace observation hole is manually input into the holographic perspective platform by artificial means; The holographic perspective platform has functions including modeling, monitoring and prediction mode switching and model updating; the modeling refers to modeling the furnace combustion process of the actual coal-fired unit, and the built model is composed of a computational fluid dynamics model and a deep learning model; the operation modes of the holographic perspective platform include a monitoring mode and a prediction mode, wherein the monitoring mode is to display the current furnace combustion process in real time according to the actual boiler operation parameters; the prediction mode is to display the combustion process of the boiler when running at the parameter set by human beings, and the two modes can be switched by artificial means; The computational fluid dynamics model adopts a partition model for calculation, specifically, the boiler furnace is divided into different regions according to the furnace cross-section temperature measurement positions, and the computational grids of the computational fluid dynamics models of the furnace temperature field, flow field and composition field are established respectively; first, the calculation starts from the area where the burner is located, and the boundary conditions are based on the actual boiler operation parameters; then, the adjacent area of the combustion area is calculated, and the calculation result of the combustion area is used as the boundary condition; and the other furnace areas are calculated in the same way; When the furnace holographic perspective platform is in the monitoring mode, the furnace holographic perspective platform model takes the current operation data and operation instructions as input, calculates the temperature field, flow field and composition field of the furnace, and then displays the calculation result in the form of an image; specifically, the results are displayed in two dimensions in the form of the cross-section, vertical section in front and back, and vertical section in left and right directions of the position specified by the operation personnel, or the above-mentioned sections are combined to realize three-dimensional display; when the platform is in the prediction mode, the model takes the set operation parameters manually input by the operation personnel in the platform as input, calculates the corresponding furnace temperature field, flow field and composition field, and displays them in the above-mentioned two-dimensional or three-dimensional way; The model updating refers to that the furnace holographic perspective platform has an automatic model updating function; when the furnace holographic perspective platform is in the monitoring mode, if the deviation between the current model calculation result and the actual cross-section temperature measurement and the temperature measurement result of the tube observation hole exceeds the set threshold, the combustion process model is automatically updated; specifically, the average variance or other type loss function of the actual cross-section temperature measurement, the temperature measurement result of the tube observation hole and the calculated value of the combustion process model is taken as the training target, and the combustion process learning model is fine-tuned.

2. A holographic see-through system for monitoring combustion in the furnace of a coal-fired boiler according to claim 1, characterized in that, The temperature field, flow field and composition field in the furnace of the actual coal-fired unit under different coal mill combination operation modes are calculated by using the computational fluid dynamics model; and based on the calculation results, the combustion process models of the temperature field, flow field and composition field in the furnace are respectively established based on deep learning; the deep learning model adopts a three-dimensional convolution network structure, the input of the combustion process model is the operation parameters of the boiler, and the output is the numerical values of the temperature field, flow field and composition field in the furnace.

3. A method of monitoring combustion in a furnace of a coal-fired boiler using the holographic see-through system of any one of claims 1-2, characterized in that, The method comprises the following steps: Step 1: obtaining the current operation data and operation instructions of the actual coal-fired unit from the DCS of the actual coal-fired unit through the holographic perspective platform, including the operation operation parameters and the furnace measurement parameters; if the furnace cross-section temperature measurement equipment output of the actual coal-fired unit is not connected to the DCS, the measurement results are directly obtained from the interface provided by the temperature measurement equipment, and at this time, the temperature measurement results of the actual coal-fired boiler furnace observation hole are manually entered into the holographic perspective platform by artificial means; Step 2: modeling the furnace combustion process of the actual coal-fired unit, the model is composed of a computational fluid dynamics model and a deep learning model; the computational fluid dynamics model adopts a partition model for calculation, and the boiler furnace is divided into different regions according to the furnace cross-section temperature measurement position, and the calculation grid of the computational fluid dynamics model of the temperature field, flow field and composition field in the furnace is established; first, the calculation starts from the region where the burner is located, and the boundary condition is based on the operation parameters of the actual boiler; then, the adjacent region of the combustion region is calculated, and the calculation result of the combustion region is taken as the boundary condition; and the other furnace regions are calculated in the same way; Step 3: the temperature field, flow field and composition field in the furnace of the actual coal-fired unit under different coal mill combination operation modes are calculated by using the computational fluid dynamics model; based on the calculation results, the combustion process models of the temperature field, flow field and composition field in the furnace are respectively established based on deep learning; the deep learning model adopts a three-dimensional convolution network structure, the input of the combustion process model is the operation parameters of the boiler, and the output is the numerical values of the temperature field, flow field and composition field in the furnace; Step 4: when the furnace holographic perspective platform is in the monitoring mode, if the deviation between the calculation results of the current model and the actual cross-section temperature measurement and the tube fire hole temperature measurement result exceeds the set threshold, the combustion process model is automatically updated, and the average variance or other type loss function of the actual cross-section temperature measurement, the tube fire hole temperature measurement result and the combustion process model calculation value is minimized as the training target, and the combustion process model is fine-tuned.

Images