Roller kiln temperature monitoring system and method

Abstract

The present application relates to the technical field of temperature control, and discloses a roller kiln temperature monitoring system and method. The method comprises the following steps: dividing temperature sensing data of a temperature sensing optical fiber and measured temperature data of a temperature sensing module into a training set and a test set according to time sequence of generation; segmenting the temperature sensing optical fiber, constructing a fiber heat conduction model, and setting a bias value for the fiber heat conduction model corresponding to each segment of the temperature sensing optical fiber; calculating error values between the temperature sensing data of each segment of the temperature sensing optical fiber and the measured temperature data of the temperature sensing module according to the fiber heat conduction model and the set bias value, and obtaining temperature error values; re-determining the bias value through a machine learning optimization algorithm; and outputting the temperature sensing data, coordinate information corresponding to the temperature sensing data and generation time according to the fiber heat conduction model and the optimized bias value. In this embodiment, a fiber heat conduction model is constructed through distributed optical fiber sensing detection technology, so as to obtain temperature field data of each position of a roller kiln lining.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of temperature control, and in particular to a roller kiln temperature monitoring system and method. BACKGROUND

[0002] As an extremely important production equipment in the ceramic process, the temperature of the kiln has a direct impact on the quality of the finished product, and the temperature distribution in the kiln is uneven. It is difficult to obtain a relatively accurate temperature field through traditional temperature measurement technology, and there is an urgent need to develop an advanced ceramic kiln temperature detection method.

[0003] The existing kiln temperature detection method is to install multiple thermocouples in multiple directions inside the kiln, and to detect the temperature inside the kiln by using the temperature reception and transmission between the thermocouples. However, due to the safety limitation of the number of openings in the furnace shell, the circumferential direction thermocouples are generally spaced 1-3m apart, so the number of temperature monitoring is limited, and there is a large monitoring blind area between the two adjacent thermocouples, which cannot achieve full-coverage monitoring of the temperature. In addition, it still uses electrical sensing, and the long-distance transmission signal attenuation is large, there are many noise sources on site, and it is greatly affected by electromagnetic interference, resulting in large errors, affecting the measurement accuracy of the kiln temperature, and unable to quickly locate the temperature abnormal point and accurately warn, which brings great safety hazards to the kiln production. SUMMARY

[0004] The purpose of the present application is to provide a roller kiln temperature monitoring system and method to solve one or more technical problems existing in the prior art, and at least to provide a beneficial choice or create conditions.

[0005] In a first aspect, a roller kiln temperature monitoring system is provided, comprising:

[0006] A temperature sensing module is configured to collect internal temperature information of the roller kiln.

[0007] A temperature sensing optical fiber is connected to the temperature sensing module, and the temperature sensing optical fiber is wrapped around the periphery of the roller kiln.

[0008] A distributed optical fiber temperature measurement module is connected to the temperature sensing optical fiber, and the distributed optical fiber temperature measurement module is configured to convert the scattered light signal of the temperature sensing optical fiber into corresponding temperature sensing data, derive an optical fiber heat conduction model according to the temperature sensing data and the measured temperature data of the temperature sensing module, and calculate the temperature field data of each position of the inner lining of the roller kiln.

[0009] Further, the temperature sensing optical fiber is wrapped around the periphery of the roller kiln in an S-shaped wrapping manner along the length direction of the roller kiln.

[0010] Further, an isolation layer is provided between the temperature sensing optical fiber and the roller kiln.

[0011] Further, the number of the temperature sensing modules is at least three, which are respectively arranged at the feeding port position, the discharging port position and the corner position of the roller kiln.

[0012] Further, the roller kiln temperature monitoring system further comprises:

[0013] The alarm module is in communication with the distributed optical fiber temperature measurement module, and the distributed optical fiber temperature measurement module is used to generate an alarm signal when the temperature field data of the inner lining of the roller kiln exceeds the preset temperature range. The alarm module is used to receive the alarm signal and give an alarm.

[0014] In a second aspect, a roller kiln temperature monitoring method is provided, which uses the roller kiln temperature monitoring system as described in the first aspect for measurement. The roller kiln temperature monitoring method comprises the following steps:

[0015] The temperature sensing data of the temperature sensing fiber and the measured temperature data of the temperature sensing module are divided into a training set and a test set according to the time sequence of generation;

[0016] The temperature sensing fiber is segmented, a fiber heat conduction model is constructed, and a bias value is set for each segment of the temperature sensing fiber corresponding to the fiber heat conduction model;

[0017] According to the fiber heat conduction model and the set bias value, the error value between the temperature sensing data of each segment of the temperature sensing fiber and the measured temperature data of the temperature sensing module is calculated to obtain a temperature error value;

[0018] Through a machine learning optimization algorithm, the bias value is re-determined by taking the optimization of the temperature error value to the minimum as the optimization goal;

[0019] According to the fiber heat conduction model and the optimized bias value, the temperature sensing data, the coordinate information corresponding to the temperature sensing data and the generation time are output.

[0020] Further, the construction of the fiber heat conduction model comprises:

[0021] The length between the position of each emission end and the temperature sensing data output position, and the length between the position of the temperature sensing module and the temperature sensing data output position are calculated according to the time difference of the scattering light signals generated by the incident end and the emission end of the temperature sensing fiber;

[0022] The light intensity ratio of different types of scattering light signals at different temperatures is calculated; wherein the scattering light signals include Stokes scattering light and anti-Stokes scattering light;

[0023] The fiber heat conduction model is determined according to the light intensity ratio of the scattering light signals at the measured temperature and the reference temperature.

[0024] Further, the fiber heat conduction model is specifically:

[0025]

[0026] wherein h is Planck's constant, c is the speed of light in vacuum, k is the Boltzmann constant, Δυ is the Raman shift, T1 is the temperature to be measured, T2 is the reference temperature, R(T1) is the light intensity ratio of the two scattered lights at the temperature to be measured, R(T2) is the light intensity ratio of the two scattered lights at the reference temperature, T 实际 is the measured temperature data of the temperature sensing optical fiber, T 偏置 is the bias value of the temperature sensing optical fiber.

[0027] Further, the bias value is re-determined by taking the optimization of the temperature error value to the minimum as an optimization target through the machine learning optimization algorithm, comprising:

[0028] randomly extracting a plurality of temperature sensing data and corresponding measured temperature data from the training set;

[0029] deriving the loss function of the optical fiber heat conduction model with respect to the bias value, and updating the temperature error value and the bias value according to the negative gradient direction of each bias value.

[0030] Further, the roller kiln temperature monitoring method further comprises:

[0031] The distributed optical fiber temperature measurement module generates an alarm signal when the temperature field data of the roller kiln lining exceeds the preset temperature range, so that the alarm module in communication with the distributed optical fiber temperature measurement module receives the alarm signal and alarms.

[0032] The beneficial effects of the present application are that the optical fiber is arranged outside the roller kiln, the internal temperature information of the roller kiln is collected at a specific position, the optical fiber heat conduction model is constructed through the distributed optical fiber sensing detection technology, the bias value of the optical fiber heat conduction model is optimized using the optimization algorithm, the temperature fed back by the optical fiber temperature sensing data is close to the internal temperature of the roller kiln, and thus the temperature field data of each position of the roller kiln lining is obtained, and the temperature conditions of each position of the roller kiln lining are accurately monitored. BRIEF DESCRIPTION OF DRAWINGS

[0033] Figure 1 is a structural schematic diagram of a model quantization training system provided by an embodiment of the present disclosure.

[0034] Figure 2 is a structural schematic diagram of a model quantization training system provided by another embodiment of the present disclosure.

[0035] Figure 3 is a flowchart of a model quantization training system provided by an embodiment of the present disclosure.

[0036] Figure 4 is Figure 3 the flowchart of step S200 in

[0037] Figure 5 is Figure 3 the flowchart of step S400 in DETAILED DESCRIPTION

[0038] In order to make the objects, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and should not be used to limit the present application.

[0039] It should be noted that although the functional modules are divided in the device schematic diagram, and the logical sequence is shown in the flowchart, in some cases, the steps shown or described can be executed in a manner different from the module division in the device or the sequence in the flowchart. The terms "first", "second", etc. in the specification and claims and the above-described drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence.

[0040] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the specification herein is for describing the embodiments of the present application only and is not intended to limit the present application.

[0041] In addition, the described features, structures or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, many specific details are provided to give a full understanding of the embodiments of the present disclosure. However, one skilled in the art will realize that the technical solutions of the present disclosure can be practiced without one or more of the specific details, or other methods, components, devices, steps, etc. can be used. In other cases, well-known methods, devices, implementations or operations are not shown or described in detail to avoid obscuring the aspects of the present disclosure.

[0042] The block diagrams shown in the drawings are only functional entities, and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0043] The flowcharts shown in the drawings are only exemplary illustrations, and do not necessarily include all contents and operations / steps, nor do they necessarily be executed in the order described. For example, some operations / steps can be further divided, and some operations / steps can be combined or partially combined, so the actual execution order can be changed according to the actual situation.

[0044] The existing kiln temperature detection method is mostly to install multiple thermocouples in multiple directions inside the kiln, to detect the temperature inside the kiln by using the temperature receiving and transmission between the thermocouples. However, due to the safety limitation of the number of openings of the furnace shell, the thermocouples in the circumferential direction are generally spaced 1-3 m apart, so the number of temperature monitoring is limited, and there is a large monitoring blind area between the two adjacent thermocouples, which cannot well realize the omnidirectional and full-coverage monitoring of the temperature. Moreover, it still uses electrical sensing, and the long-distance transmission signal attenuation is large, there are many noise sources on site, and it is greatly interfered by electromagnetic interference, resulting in large error, affecting the measurement accuracy of the kiln temperature, and unable to quickly locate the temperature abnormal point and accurately warn, which brings great safety hazard to the kiln production.

[0045] Based on this, the present embodiment proposes a roller kiln temperature monitoring system and method, which sets optical fibers outside the roller kiln and collects internal temperature information of the roller kiln at specific positions, constructs a fiber heat conduction model through a distributed optical fiber sensing detection technology, and calculates the temperature field data of each position of the roller kiln lining to accurately monitor the temperature of each position of the roller kiln lining.

[0046] The "roller kiln" referred to herein refers to an archless kiln, which is a tunnel kiln with a long and narrow cross section, composed of rollers arranged in parallel across the cross section of the kiln working channel to form a "roller way". The products are placed on the roller way and transported into the kiln with the rotation of the rollers to complete the firing process in the kiln.

[0047] Referring to Figures 1 to 2 The roller kiln temperature monitoring system provided by the various exemplary embodiments of the present application comprises a temperature sensing module 100, a temperature sensing optical fiber 200, and a distributed optical fiber temperature measurement module 300.

[0048] The temperature sensing module 100 is used to collect internal temperature information of the roller kiln; the temperature sensing optical fiber 200 is connected with the temperature sensing module 100, and the temperature sensing optical fiber 200 is wrapped around the periphery of the roller kiln; the distributed optical fiber temperature measurement module 300 is connected with the temperature sensing optical fiber 200, and the distributed optical fiber temperature measurement module 300 is used to convert the scattered light signal of the temperature sensing optical fiber 200 into corresponding temperature sensing data, to derive a fiber heat conduction model according to the temperature sensing data and the measured temperature data of the temperature sensing module 100, and to calculate the temperature field data of each position of the roller kiln lining.

[0049] The temperature sensing module 100 can be a thermocouple sensor, a thermistor sensor, a resistance temperature detector or an IC temperature sensor, which converts the internal temperature information of the roller kiln into corresponding sensing signals. The temperature sensing principle of the temperature sensing fiber 200 can be that the fiber core diameter and the refractive index of the temperature sensing fiber 200 change with temperature, so that the light propagating in the temperature sensing fiber 200 scatters outward due to uneven routes, resulting in changes in light amplitude, or the length, refractive index and fiber core diameter of the temperature sensing fiber 200 change with temperature, so that the light propagating in the temperature sensing fiber 200 changes in phase.

[0050] The roller kiln temperature monitoring system of the exemplary embodiments of the present application receives the temperature sensing data returned by the temperature sensing fiber 200 in real time and the measured temperature data of the temperature sensing module 100 through the distributed fiber temperature measurement module 300, derives a fiber heat conduction model, uses the fiber heat conduction model to locate the real-time temperature information of each position of the fiber, thereby obtaining the temperature field data of each position of the inner lining of the roller kiln, and determines whether to alarm according to the preset temperature threshold value through the temperature measurement software. When the measured temperature corresponding to the temperature sensing data is higher than the preset temperature threshold value, an alarm information will be sent, and the temperature of each position of the inner lining of the roller kiln is monitored.

[0051] In an embodiment, as shown in Figure 1 and Figure 2 , the temperature sensing fiber 200 is wrapped around the periphery of the roller kiln in an S-shaped wrapping manner along the length direction of the roller kiln. The temperature sensing fiber 200 is wrapped around the periphery of the roller kiln in an S-shaped wrapping manner along the length direction of the roller kiln, and the end of the temperature sensing fiber 200 is connected to the distributed fiber temperature measurement module 300, which can quickly, continuously and in real time detect the temperature field signals of all directions of the inner lining of the roller kiln.

[0052] In an embodiment, as shown in Figure 1 and Figure 2 , an isolation layer 400 is arranged between the temperature sensing fiber 200 and the roller kiln. The isolation layer 400 can be wrapped around the periphery of the temperature sensing fiber 200, or can be covered on the outer wall surface of the roller kiln. The isolation layer 400 is used to protect the temperature sensing fiber 200, so as to avoid direct contact between the temperature sensing fiber 200 and the roller kiln. The temperature sensing fiber 200 indirectly measures the internal temperature of the roller kiln, that is, the temperature corresponding to the temperature sensing data of the temperature sensing fiber 200 is not the same as the internal temperature of the roller kiln. When the fiber heat conduction model is constructed, the distributed fiber temperature measurement module 300 needs to introduce a bias value to compensate for the deviation between the temperature sensing data of the temperature sensing fiber 200 and the internal temperature of the roller kiln.

[0053] In an embodiment, the number of temperature sensing modules 100 is at least three, which are respectively arranged at the inlet position, the outlet position and the corner position of the roller kiln.

[0054] In an embodiment, as shown in Figure 2As shown, the roller kiln temperature monitoring system further comprises an alarm module 500, the alarm module 500 is in communication with the distributed optical fiber temperature measurement module 300, the distributed optical fiber temperature measurement module 300 is used to generate an alarm signal when the roller kiln lining temperature field data exceeds the preset temperature range, and the alarm module 500 is used to receive the alarm signal and alarm. The alarm module 500 can be an alarm or an alarm lamp. For example, when the distributed optical fiber temperature measurement module 300 determines that the i-th measured temperature data of the n-th temperature sensing module 100 is greater than the preset temperature threshold, or the temperature corresponding to the i-th temperature sensing data of the n-th temperature sensing part of the temperature sensing optical fiber 200 is greater than the preset temperature threshold, the distributed optical fiber temperature measurement module 300 automatically outputs an alarm signal, and the alarm module 500 alarms, and sends a request to stop or slow down the fuel into the roller kiln to the workstation until the measured temperature data of the temperature sensing module 100 and the temperature sensing data of the temperature sensing optical fiber 200 return to below the preset temperature threshold.

[0055] The roller kiln temperature monitoring system provided in the embodiment sets optical fibers outside the roller kiln and collects internal temperature information of the roller kiln at specific positions, constructs an optical fiber heat conduction model through a distributed optical fiber sensing detection technology, optimizes the bias value of the optical fiber heat conduction model using an optimization algorithm, makes the temperature fed back by the temperature sensing data of the optical fiber close to the internal temperature of the roller kiln, and thus obtains the temperature field data of each position of the roller kiln lining and accurately monitors the temperature conditions of each position of the roller kiln lining.

[0056] Based on the same concept, the embodiment of the present application also provides a roller kiln temperature monitoring method.

[0057] Reference Figure 3 According to the roller kiln temperature monitoring method provided by the example embodiment of the present disclosure, the roller kiln temperature monitoring method is measured by using the roller kiln temperature monitoring system described above, and the roller kiln temperature monitoring method comprises but is not limited to steps S100 to S500.

[0058] In step S100, the temperature sensing data of the temperature sensing optical fiber and the measured temperature data of the temperature sensing module are divided into a training set and a test set in chronological order;

[0059] In step S200, the temperature sensing optical fiber is segmented, an optical fiber heat conduction model is constructed, and a bias value is set for each segment of the temperature sensing optical fiber corresponding to the optical fiber heat conduction model;

[0060] In step S300, according to the optical fiber heat conduction model and the set bias value, the error value between the temperature sensing data of each segment of the temperature sensing optical fiber and the measured temperature data of the temperature sensing module is calculated, and a temperature error value is obtained;

[0061] In step S400, a machine learning optimization algorithm is used to optimize the temperature error value to the minimum as an optimization target, and the bias value is re-determined.

[0062] In step S500, the temperature sensing data, the coordinate information corresponding to the temperature sensing data and the generation time are output according to the optical fiber heat conduction model and the optimized bias value.

[0063] In step S100 of some embodiments, the temperature sensing data of the temperature sensing fiber and the measured temperature data of the temperature sensing module are divided into a training set and a test set. Specifically, the temperature sensing data of the temperature sensing fiber and the measured temperature data of the temperature sensing module corresponding to the setting position of the temperature sensing module are matched. The temperature sensing data and the measured temperature data generated at the same time are matched as a group of temperature data. With the passage of time, each temperature sensing module and the corresponding position of the temperature sensing fiber generate a plurality of groups of temperature data. These temperature data are proportionally divided. A part of the proportion of the temperature data is divided into the training set, and another part of the proportion of the temperature data is divided into the test set.

[0064] In step S200 of some embodiments, the temperature sensing fiber is segmented. The temperature sensing fiber can be segmented according to the setting position of the temperature sensing module. For example, the number of temperature sensing modules is three, which are respectively configured at the feeding port position, the discharging port position and the corner position of the roller kiln. The temperature sensing fiber is segmented according to the setting position of the temperature sensing module. The temperature sensing fiber between two temperature sensing modules is divided into a segment. After the temperature sensing fiber is segmented, the optical fiber heat conduction model is constructed for each segmented temperature sensing fiber. The bias value is set for the optical fiber heat conduction model corresponding to each segmented temperature sensing fiber to preliminarily form the optical fiber heat conduction model. The bias value can be set artificially according to past experience. The error between the temperature sensing data of the temperature sensing fiber and the measured temperature data of the temperature sensing module is as little as possible.

[0065] In step S300 of some embodiments, the error value between the temperature sensing data of each segmented temperature sensing fiber and the measured temperature data of the temperature sensing module is calculated according to the optical fiber heat conduction model and the set bias value to obtain the temperature error value. Specifically, the error value between the temperature sensing data and the measured temperature data is calculated according to the optical fiber heat conduction model initially set in step S200 and the plurality of groups of temperature data obtained in step S100. The temperature sensing data is converted from the scattered light signal of the temperature sensing fiber. The deviation between the scattered light signal of the temperature sensing fiber indirectly collected from the roller kiln temperature and the actual temperature inside the roller kiln is compensated by the set bias value.

[0066] In step S400 of some embodiments, the bias value is re-determined by a machine learning optimization algorithm to optimize the temperature error value to a minimum as an optimization objective. Specifically, after the temperature error value calculated in step S300, the bias value is corrected by the machine learning optimization algorithm, so that the temperature error value gradually decreases after each correction of the bias value, so that the compensation of the bias value to the temperature sensing data gradually approaches the actual temperature inside the roller kiln.

[0067] In step S500 of some embodiments, according to the optical fiber heat conduction model and the optimized bias value, the temperature sensing data and the coordinate information and the generation time corresponding to the temperature sensing data are output. Specifically, the optical fiber heat conduction model optimized by the machine learning optimization algorithm calculates the temperature sensing data according to the scattered light signals generated by the temperature sensing fiber at each position, and the calculated temperature sensing data and the corresponding coordinate information and generation time are output to an external workstation or uploaded to a cloud server for display. The coordinate information corresponding to the temperature sensing data can be determined according to the time difference between the scattered light signals generated by the incident end and the emitting end of the temperature sensing fiber.

[0068] The roller kiln temperature monitoring method provided in the embodiment sets optical fibers outside the roller kiln and collects internal temperature information of the roller kiln at specific positions, constructs an optical fiber heat conduction model by using a distributed optical fiber sensing detection technology, optimizes the bias value of the optical fiber heat conduction model by using an optimization algorithm, makes the temperature feedback by the optical fiber temperature sensing data close to the internal temperature of the roller kiln, and thus obtains the temperature field data of each position of the roller kiln lining and accurately monitors the temperature conditions of each position of the roller kiln lining.

[0069] In some embodiments, as shown in FIG. 2, Figure 4 Step S200 specifically includes but is not limited to steps S210 to S230.

[0070] Step S210 calculates the length between each emitting end and the temperature sensing data output position, and the length between the position of the temperature sensing module and the temperature sensing data output position according to the time difference between the scattered light signals generated by the incident end and the emitting end of the temperature sensing fiber;

[0071] Step S220 calculates the light intensity ratio of different types of scattered light signals at different temperatures; wherein the scattered light signals include Stokes scattered light and anti-Stokes scattered light;

[0072] Step S230 determines the optical fiber heat conduction model according to the light intensity ratio of the scattered light signals at the measured temperature and the reference temperature.

[0073] In step S210 of some embodiments, the length between the position of each emission end and the position of the temperature sensing data output, and the length between the position of the temperature sensing module and the position of the temperature sensing data output are calculated according to the time difference of the scattered light signals generated by the temperature sensing fiber incident end and the emission end. Specifically, the temperature sensing fiber has a plurality of emission ends for emitting scattered light, and the temperature sensing module is correspondingly arranged at the positions of the plurality of emission points. The temperature sensing fiber can be precisely positioned along the temperature measurement points by the optical time domain reflection principle (OTDR), and the time difference between the incident end and the scattered light is measured, so as to calculate the distance between the emission end of the scattered light and the incident end. The specific calculation formula is as follows:

[0074] l = ct / 2n;

[0075] wherein, l is the distance between the emission end and the incident end, c is the speed of light in vacuum, t is the time difference between the incident end and the scattered light, and n is the refractive index of the temperature sensing fiber.

[0076] In step S220 of some embodiments, the light intensity ratio of different types of scattered light signals at different temperatures is calculated. Specifically, the scattered light signal reflected to the incident end contains Stokes scattered light (Stokes) and Anti-Stokes scattered light (Anti-Stokes), and the Anti-Stokes scattered light is sensitive to temperature and is signal light, while the Stokes scattered light is not sensitive to temperature and is reference light.

[0077] The light intensity ratio of the two scattered light signals at the measured temperature is as follows:

[0078]

[0079] The light intensity ratio of the two scattered light signals at the reference temperature is as follows:

[0080]

[0081] wherein, K s and K as are coefficients related to the cross section, V s and V as are photons generated by the interaction of Stokes scattered light and Anti-Stokes scattered light with the medium molecules in the temperature sensing fiber, α s and α as are the transmission losses of the Stokes scattered light and the Anti-Stokes scattered light in the temperature sensing fiber, h is the Planck constant, k is the Boltzmann constant, Δυ is the Raman shift amount, T1 is the measured temperature of the temperature sensing fiber, T2 is the reference temperature measured by the temperature sensing fiber, R(T1) is the light intensity ratio of the two scattered light signals at the measured temperature, and R(T2) is the light intensity ratio of the two scattered light signals at the reference temperature.

[0082] The ratio of the light intensity at the to-be-measured temperature and the ratio of the light intensity at the reference temperature is divided, and a theoretical temperature model of the temperature sensing optical fiber is obtained, and the theoretical temperature model of the temperature sensing optical fiber is specifically:

[0083]

[0084] T 实际 = T 理论 + T 偏置 = T1+ T 偏置 The optical fiber heat conduction model is obtained, and the optical fiber heat conduction model is specifically:

[0085]

[0086] Wherein, T 实际 is the measured temperature sensing data of the temperature sensing optical fiber, and T 偏置 is the bias value of the temperature sensing optical fiber.

[0087] In some embodiments, as shown in Figure 5 , the step S400 specifically includes but is not limited to steps S410 to S420.

[0088] Step S410, a plurality of temperature sensing data and corresponding measured temperature data are randomly extracted from the training set;

[0089] Step S420, the minimum value of the temperature error value obtained by using the gradient descent method is updated according to the negative gradient direction of each bias value.

[0090] In step S410 of some embodiments, a plurality of temperature sensing data and corresponding measured temperature data are randomly extracted from the training set. Specifically, a batch of temperature sensing data samples {x1,…,xm} with a capacity of m and a batch of corresponding measured temperature data samples {y1,…,ym} with a capacity of m can be randomly extracted from the training set. m . m .

[0091] In step S420 of some embodiments, the minimum value of the temperature error value obtained using the gradient descent method is updated in the negative gradient direction of each bias value. Specifically, the minimum value of the temperature error value obtained using the gradient descent method is a global optimal solution, which can be selected by using a suitable loss function of the fiber heat conduction model constructed in step S200, and then the loss function of the fiber heat conduction model is differentiated with respect to the bias value, and then the temperature error value and the bias value are updated in the negative gradient direction of each bias value. Since step S410 is random extraction, the gradient obtained is inevitably error, so the extraction and gradient descent method for obtaining the global optimal solution is repeated until the minimum value of the temperature error value is stable at the preset threshold value, or the number of repetitions reaches the upper limit of the repetition.

[0092] In some embodiments, the roller kiln temperature monitoring method further comprises: when the distributed optical fiber temperature measurement module generates an alarm signal when the roller kiln lining temperature field data exceeds the preset temperature range, the alarm module in communication with the distributed optical fiber temperature measurement module receives the alarm signal and alarms.

[0093] The preferred embodiments of the present application are described above with reference to the accompanying drawings, and are not limited to the scope of the present application. The above-mentioned embodiment numbers of the present application are only for description, and do not represent the advantages and disadvantages of the embodiments. In addition, although the logical order is shown in the flowchart, in some cases, the steps shown or described can be performed in an order different from that shown here.

[0094] Those skilled in the art can have various modifications without departing from the scope and essence of the present application, such as using the features of one embodiment in another embodiment to obtain another embodiment. The preferred embodiments of the above-mentioned embodiments of the present disclosure are described with reference to the accompanying drawings, and are not limited to the scope of the embodiments of the present disclosure. Any modification, equivalent replacement and improvement made by those skilled in the art within the scope and essence of the embodiments of the present disclosure shall be within the scope of the embodiments of the present disclosure.

Claims

1. A method of temperature monitoring of a roller kiln, characterized in that, The measurement is performed using a roller kiln temperature monitoring system; The roller kiln temperature monitoring system comprises: a temperature sensing module for collecting internal temperature information of the roller kiln; a temperature sensing fiber connected to the temperature sensing module, the temperature sensing fiber being wrapped around the periphery of the roller kiln; a distributed optical fiber temperature measurement module connected to the temperature sensing fiber, the distributed optical fiber temperature measurement module being configured to convert the scattered light signals of the temperature sensing fiber into corresponding temperature sensing data, derive a fiber heat conduction model based on the temperature sensing data and the measured temperature data of the temperature sensing module, and calculate the temperature field data of each position of the inner lining of the roller kiln. The roller kiln temperature monitoring method comprises: dividing the temperature sensing data of the temperature sensing fiber and the measured temperature data of the temperature sensing module into a training set and a test set according to the time sequence of generation; segmenting the temperature sensing fiber, constructing a fiber heat conduction model, and setting a bias value for each segment of the temperature sensing fiber corresponding to the fiber heat conduction model; calculating the error value between the temperature sensing data of each segment of the temperature sensing fiber and the measured temperature data of the temperature sensing module according to the fiber heat conduction model and the set bias value, and obtaining the temperature error value; re-determining the bias value by using a machine learning optimization algorithm to optimize the temperature error value to a minimum as an optimization goal; outputting the temperature sensing data, the coordinate information corresponding to the temperature sensing data, and the generation time according to the fiber heat conduction model and the optimized bias value. The construction of the fiber heat conduction model comprises: calculating the length between each emitting end position and the temperature sensing data output position, and the length between the temperature sensing module position and the temperature sensing data output position according to the time difference of the scattered light signals generated by the incident end and the emitting end of the temperature sensing fiber; calculating the light intensity ratio of different types of scattered light signals at different temperatures; wherein the scattered light signals include Stokes scattered light and anti-Stokes scattered light; determining the fiber heat conduction model according to the light intensity ratio of the scattered light signals at the measured temperature and the reference temperature; The fiber heat conduction model is specifically: wherein h is Planck's constant, c is the speed of light in vacuum, k is the Boltzmann constant, is the Raman shift amount, T1 is the temperature to be measured, T2 is the reference temperature, R(T1) is the light intensity ratio of the two scattered lights at the temperature to be measured, R(T2) is the light intensity ratio of the two scattered lights at the reference temperature, T 实际 is the measured temperature sensing data of the temperature sensing optical fiber, T 偏置 is the bias value of the temperature sensing optical fiber.

2. The roller kiln temperature monitoring method according to claim 1, characterized in that, The temperature sensing fiber is wrapped around the periphery of the roller kiln in an S-shaped manner along the length direction of the roller kiln.

3. The roller kiln temperature monitoring method according to claim 1, characterized in that, An isolation layer is provided between the temperature sensing fiber and the roller kiln.

4. The roller kiln temperature monitoring method according to claim 1, characterized in that, The number of temperature sensing modules is at least three, which are respectively arranged at the inlet position, outlet position and corner position of the roller kiln.

5. The roller kiln temperature monitoring method according to claim 1, characterized in that, Further comprising: an alarm module in communication with the distributed optical fiber temperature measurement module, the distributed optical fiber temperature measurement module being configured to generate an alarm signal when the temperature field data of the inner lining of the roller kiln exceeds the preset temperature range, and the alarm module being configured to receive the alarm signal and perform an alarm.

6. The roller kiln temperature monitoring method according to claim 1, characterized in that, The re-determination of the bias value by using a machine learning optimization algorithm to optimize the temperature error value to a minimum as an optimization goal comprises: randomly extracting a number of temperature sensing data and corresponding measured temperature data from the training set; calculating the partial derivative of the loss function of the fiber heat conduction model with respect to the bias value, and updating the temperature error value and the bias value according to the negative gradient direction of each bias value.

7. The roller kiln temperature monitoring method according to claim 1, characterized in that, The roller kiln temperature monitoring method further comprises: The distributed optical fiber temperature measurement module generates an alarm signal when the temperature field data of the inner lining of the roller kiln exceeds the preset temperature range, so that the alarm module in communication with the distributed optical fiber temperature measurement module receives the alarm signal and performs an alarm.

Images