Towards prediction of geomaterial properties post-processing via high-throughput chemical and spatial characterization

Abstract

Methods, systems, and apparatus, including computer programs encoded on a storage device, for determining characteristics of particles are disclosed. A system includes a conveyor belt, wherein a first section of the conveyor belt comprises one or more mechanisms arranged to create environmental conditions favorable to sensor reading. The system includes a sensor rig located proximate to the first section of the conveyor belt comprising a plurality of different types of sensors. The system includes a controller configured to: obtain measurement data from the plurality of different types of sensors; apply the measurement data to a machine learning model configured to determine surface chemical characteristics of particles given measurement data; and output the surface chemical characteristics of particles from the machine learning model to one or more components along the conveyor belt.

Description

TOWARDS PREDICTION OF GEOMATERIAL PROPERTIES POST PROCESSING VIA HIGH-THROUGHPUT CHEMICAL AND SPATIAL CHARACTERIZATIONBACKGROUND[1] Mining geo-materials for different geo-material processing procedures to produce a particular product is typically composed of particles. The quality of the product and efficiency of manufacturing the product depends on the characteristics of the aggregate and how the characteristics affect the geo-material processing procedure. For example, different concrete mixes can result in concrete of different strength, or different amounts of carbon dioxide emitted during the manufacturing process. However, the aggregate can range in physical composition and / or chemical composition, which can alter the manufacturing process by resulting in inefficiencies or latency in processing the aggregate. Conventional methods for monitoring aggregate may provide limited information about relevant characteristics of the particles of the aggregate, leading to suboptimal downstream processing decisions.SUMMARY[2] This specification relates generally to characterizing particles on a conveyor belt.[3] In general, the disclosure relates to a system for determining characteristics of particles conveyed through a particle processing system, e g., on a conveyor belt. For example, systems and processes described herein determine characteristics of particles located on a conveyor belt, in order to provide relevant and up-to-date information for downstream processing systems. By determining characteristics such as surface area, volume, and composition of particles traveling on a conveyor belt, a downstream processing system can determine an optimal concrete mix for the current particles on the conveyor belt, an adjusted direction for mining to obtain particles with certain characteristics, or a setting for a tool or piece of machinery.L4J A material processing system can use data from a combination of sensors to determine characteristics of particles such as surface area, volume, composition, shape, water content, specific surface area, surface roughness, etc. Sensors can include, for example, cameras, infrared cameras, hyperspectral imaging cameras, lasers, UV lasers, X-rayfluorescence (XRF) sensors, magnetic resonance imaging (MRI) sensors, laser displacement sensors, and / or radiation sensors. The sensors can obtain measurement data for a batch of the particles. The data can be applied as input to a machine learning model that is trained to determine characteristics of a batch of particles given measurement data. The machinelearning model can be trained, for example, using measurement data of 3D-printed particles or particles of known characteristics.[5] The material processing system can include features to improve the accuracy of predictions from images taken in locations that are different from those depicted in images used to train the model. For example, the conditions of each conveyor belt system may differ in lighting, background characteristics, etc. Such features can include sampling a batch from the conveyor belt, for example, at regular intervals of time. The sampled batch can be directed into a separate section of the conveyor belt system with mechanisms arranged to create environmental conditions favorable to sensor reading. For example, the separate section can include similar lighting and other conditions as the images the model was trained on. After the sensors collect data, the sampled batch can be redirected back to the conveyor belt.[6] In some instances, the techniques can be used simultaneously. For example, a conveyor belt system can include multiple sensors aiming at the same location of the conveyor belt that provide data to a model, and the location of the conveyor belt has the same conditions as the images the model was trained on.[7] In general, innovative aspects of the subject matter described in this specification can be embodied in a system, non-transitory computer readable medium, and methods that include the actions of obtaining, by one or more processors, measurement data from a plurality of different types of sensors arranged relative to a particular section of a conveyor belt to obtain measurements of particles conveyed along the particular section of the conveyor belt. The particular section of the conveyor belt includes one or more mechanisms arranged to create environmental conditions favorable to sensor reading. The method includes applying, by one or more processors, the measurement data to a machine learning model trained to determine surface chemical characteristics of particles given measurement data. The method includes outputting, by the one or more processors, relevant characteristics of particles from the machine learning model to one or more components along the conveyor belt.[8] Other implementations of this aspect include corresponding systems, apparatus, and computer programs, configured to perform the actions of the methods, encoded on computer storage devices.[9] One aspect can be embodied in a system for determining characteristics of particles that includes: a conveyor belt, where a first section of the conveyor belt includes one or more mechanisms arranged to create environmental conditions favorable to sensor reading;a sensor rig located proximate to the first section of the conveyor belt comprising a plurality of different types of sensors arranged to obtain measurements of particles conveyed by the first section of the conveyor belt; and a controller configured to: obtain measurement data from the plurality of different types of sensors; apply the measurement data to a machine learning model trained to determine surface chemical characteristics of particles given measurement data; and output relevant characteristics of particles from the machine learning model to one or more components along the conveyor belt.

[0010] These and other implementations can each optionally include one or more of the following features.

[0011] In some implementations, the surface chemical characteristics of particles include any one of: an oxidation rate, a reactivity rate, a chemical composition ratio, or a moisture evaporation rate.

[0012] In some implementations, the machine learning model is trained on training examples comprising measurement data of particles of known surface chemical characteristics.

[0013] In some implementations, the measurement data includes images of the particles, and where the images of the particles are thermal images, infrared images, hyper-spectral images, or a combination thereof.

[0014] In some implementations, the machine learning model determines the surface chemical characteristics of the particles by identifying regions of pixels of the images corresponding to the particles.

[0015] In some implementations, determining the surface chemical characteristics of the particles includes identifying one or more chemicals on the surface of the particles based on identifying the regions of pixels of the images; and generating a chemical composition ratio for the particles, where the chemical composition includes a percentage content of each of the one or more chemicals.

[0016] In some implementations, the plurality of different ty pes of sensors are positioned along the conveyor belt to point at different regions of the conveyor belt, and the controller is further configured to synchronize measurement data from the sensors.

[0017] In some implementations, the plurality of sensors includes optical sensors, infrared sensors, or a combination thereof.

[0018] In some implementations, at least one of the plurality of different types of sensors is positioned to capture measurement data of a top surface of the particle, and where thecontroller is configured to determine a surface area for each of the particles based on the measurement data of the top surface of the particle.

[0019] In some implementations, the sensor rig includes a spray system configured to wet the particles to obtain wetted particles prior to obtaining the measurement data of the wetted particles from an infrared sensor of the plurality of sensors. .

[0020] In some implementations, the system determines a moisture evaporation rate at the surface of the wetted particles based on obtaining the measurement data of the wetted particles.

[0021] In some implementations, the controller is configured to apply the measurement data to a second machine learning model trained to determine physical characteristics of particles given the measurement data.

[0022] In some implementations, the physical characteristics of particles include any one of: surface area, volume, composition, shape, water content, specific surface area, or surface roughness.

[0023] In some implementations, the second machine learning model is trained on training examples comprising measurement data of particles of known physical characteristics

[0024] In some implementations, different relevant characteristics are outputted to different components along the conveyor belt.

[0025] In some implementations, the controller is further configured to adjust control settings of a chemical processing subsystem based at least in part on the surface chemical characteristics, the physical characteristics, or both.

[0026] In some implementations, the controller is further configured to determine adjustments to control settings of a particle processing system based at least in part on the surface chemical characteristics, the physical characteristics, or both.

[0027] In some implementations, the controller is further configured to calculate a reactivity of the particles based at least in part on the surface chemical characteristics, the physical characteristics, or both; and determine adjustments to the control settings based on the reactivity.

[0028] In some implementations, the adjustments to the control settings include: an adjustment to a flow rate of the particle processing system, a temperature of the particle processing system, or a combination thereof.

[0029] The subject matter described in this specification can be implemented in particular embodiments so as to realize one or more of the following advantages.

[0030] The system can ensure that downstream processing decisions for processing particles are more informed. For example, by reducing emissions to perform certain processes, such as carbonation, the system can reduce carbon dioxide emissions. In addition, the surface chemical properties of particles are inconsistent due to the large variation in ingredient material (e.g., particles) and processing. For example, depending on the surface chemical characteristics of the particles and the physical characteristics of the particles, the post-processing procedures can differ in the amounts of other materials, such as water, that are needed. This material inconsistency requires large safety margins for a given performance level and results in material overuse. Advances in particle preparation that can optimize the use of locally available materials to maximize post-processing performance while minimizing cost with added ingredients are desirable.

[0031] The system can determine surface chemical characteristics of particles traveling on a conveyor belt using multiple types of sensors. Combining measurement data from multiple types of sensors can result in more accurate determination of the surface chemical characteristics, and the machine learning model can more accurately determine characteristics of the particles in order to provide the characteristics to a particle processing system, such a blast furnace system, a carbonation system, or a densification system.

[0032] The details of one or more embodiments of the subject matter of this specification are set forth in the accompanying drawings and the description below. Other features, aspects, and advantages of the subject matter will become apparent from the description, the drawings, and the claims.BRIEF DESCRIPTION OF THE DRAWINGS

[0033] FIG. 1 is a block diagram of the example system for determining characteristics of particles on a conveyor belt.

[0034] FIG. 2 is a view of an example section on a conveyor belt.

[0035] FIGs. 3A and 3B are exemplary' particle processing systems for processing outputs of the conveyor belt.

[0036] FIG. 4 is a block diagram of an exemplary control system for the system of FIG. 1.

[0037] FIG. 5 is a flow' chart of an example process for determining surface chemical characteristics of particles.

[0038] FIG. 6 depicts a schematic diagram of a computer system that may be applied to any of the computer-implemented methods and other techniques described herein.

[0039] Like reference numbers and designations in the various drawings indicate like elements.DETAILED DESCRIPTION

[0040] FIG. 1 is a diagram of the example system 100 for determining surface chemical characteristics of particles 130 in order to provide characteristics 134 to one or more example particle processing systems. The particles 130 can be, for example, rocks. The system 100 includes a pre-processing subsystem 102, a conveyor belt 104, a controller 106, and an example particle processing systems 108. Some or all of the functions of the system 100 can be implemented in electronic circuitry', e.g., by individual computer systems (e.g., servers), processors, microcontrollers, a field programmable gate array (FPGA), or an application specific integrated circuit (ASIC). For example, the pre-processing subsystem 102, controller 106, and example particle processing systems 108 can be provided as one or more computer executable software modules, hardware modules, or a combination thereof. For example, one or more of the pre-processing subsystem 102, controller 106, or one or more example particle processing systems 108 can be implemented as blocks of software code with instructions that cause one or more processors of the control system 200 to execute operations described herein. In addition or alternatively, one or more of the pre-processing subsystem 102, controller 106, or example particle processing systems 108 can be implemented in electronic circuitry such as, e.g., programmable logic circuits, field programmable logic arrays (FPGA), or application specific integrated circuits (ASIC).

[0041] The system 100 can determine surface chemical characteristics of particles or materials being conveyed on the conveyor belt 104. In particular, the surface chemical characteristics of the particles can be an oxidation rate, a reactivity rate, a chemical composition ratio, a moisture evaporation rate, or a combination thereof. The materials may be included in a material mixture or final product. The system 100 can be part of a material preparation system. For example, a material preparation system can use the system 100 to adaptively adjust proportions of materials in the material mixture. In this case, the material preparation system uses the system 100 to adjust proportions of materials for smelting (e.g., iron ore smelting) using a blast furnace. In particular, the system can adjust characteristics of iron ore particles for the smelting process, as described in further detail below with reference to FIG. 3.

[0042] The system 100 can adjust the proportions based on surface chemical characteristics determined by the system 100 to achieve a set of target properties of the mixture, for example. Target properties of the final product can include, for example, surfacechemical characteristics, and in some examples, physical characteristics. In some examples, the characteristics 134 can further include structural properties, electrical properties, thermal properties, chemical properties, or the like, of a final product.

[0043] The system includes the pre-processing subsystem 102. The pre-processing subsy stem 102 can be a physical processing subsystem 110, a chemical processing subsystem 112, or both. In some other examples, the physical processing subsystem 110 can include a crusher. The setting for the crusher may depend on a configured particle size of the particles 130, for example.

[0044] In some examples, the physical processing subsystem 110 can be a pelletization system that processes materials to generate the particles 130. The pelletization system can crush iron ores to particular physical shapes and / or geometries. In some examples, the chemical processing subsystem 112 can be a subsystem that combines one or more chemicals to chemically generate the composition of the particles 130. In general, the pre-processing subsy stem 102 can generate the particles 130 and generate surface chemical properties and structural properties of the particles 130 by adaptively adjusting control parameters of the respective physical processing subsystem 110, chemical processing subsystem 112, or both. The pre-processing subsystem 102 can provide the particles 130 to the conveyor belt 104.

[0045] The conveyor belt 104 is a conveyor belt system with at least one section that includes one or more mechanisms to create environmental conditions for sensor reading. Environmental conditions favorable to sensor reading can include favorable or ideal lighting or illumination conditions or temperature conditions.

[0046] The conveyor belt 104 includes a lighting system 114, a sensor rig 116, and, in some implementations, a sprayer system 118. The lighting system 114 includes multiple lighting devices arranged to direct light at a specific angle with respect to at least one of the sensors, as described in further detail below with reference to FIG. 2.

[0047] The sensor rig 116 is located proximate to the first section of the conveyor belt 104. The sensor rig 116 includes multiple sensors for capturing measurement data 132. The sensors can include infrared (IR) cameras. UV lasers, X-ray fluorescence (XRF) sensors, laser displacement sensors, or magnetic resonance imaging (MRI) sensors. Other sensors can include stereo cameras. In particular, the IR cameras can be near infrared (NIR) or short wave infrared (SWIR) sensors that are configured to capture infrared images of the particles 130. The SWIR sensors can capture the infrared images using visible light, non-visible light, or both. In some examples, the optical sensors or infrared sensors can be hyperspectral sensors. The sensors can be arranged on the sensor rig to obtain measurements of particles conveyedon the first section of the conveyor belt, as described in further detail below with reference toFIG. 2.

[0048] The sensors can be sensors of different types and / or sensors arranged at different positions or angles relative to the batch of particles for obtaining the measurement data 132. In some examples, the different types of sensors are positioned to capture measurement data 132 of a top surface of the particle.

[0049] The sensors can obtain different types of measurement data 132. The system performs sensor fusion by obtaining different types of sensor data to generate measurement data 132. For example, the types of measurement data can include visible light images, IR images, UV images, XRF data, hyper-spectral (e.g., multi-spectral) images, or MRI images. The sensors 202 can send the measurement data 132 to the controller 106 for fusion of the data. For example, the sensors 202 can send the measurement data 132 to the controller 106 upon receiving a request from the controller 106.

[0050] In some examples, the sprayer system 118 is configured to wet the particles 130 to obtain wetted particles. In particular, the sprayer system 118 includes one or more sprayers along the conveyor that are configured to output liquid, such as water, to wet the particles 130, and the sensor rig 116 is configured to send measurement data 132 of the top surface of the wetted particle, as described in further detail below with reference to FIG. 2. For example, the sprayer system 118 can wet the particles, and a corresponding sensor of the system (e.g., a SWIR sensor) can capture measurement data 132 including infrared images or hyper-spectral images of the wetted particles to determine a surface evaporation rate. As such, the sprayer system 118 and the one or more sensors of the system can function in conjunction to generate the measurement data 132.

[0051] The controller 106 can apply the measurement data 132 to surface chemical characteristic algorithms 120, physical characteristic machine learning algorithms 122, or both. For example, the controller 106 can provide the measurement data 132 as input to the surface chemical characteristic algorithms 120 and determine surface chemical characteristics of the particles 130. In some examples, the surface chemical characteristics algorithms 120, the physical characteristic algorithms 122. or both include a machine learning model trained to determine the particular characteristics of the particles 130. In one example, the controller 106 can provide the measurement data 132 as input to the physical characteristic machine learning model 122 trained to determine physical characteristics of the particles 130.

[0052] In some implementations, the controller 106 can perform pre-processing of the measurement data 132. For example, the controller 106 can rescale, flip, or adjust thebrightness or other properties of a visible light image. The controller 106 can also derive features from a visible light image and include the features in the input to the surface chemical characteristic algorithms 120, the physical characteristic algorithms 122, or both. For example, the controller 106 can extract color histograms or other features that describe the visible light image. In some examples, the controller 106 is configured to determine a surface area for each of the particles based on the measurement data 132 of the top surface of the particle.

[0053] The surface chemical characteristic algorithms 120 can determine surface chemical characteristics of particles given measurement data 132. For example, a machine learning model of the surface chemical characteristic algorithms 120 can be trained to determine surface chemical characteristics of particles given multiple types of measurement data. Some characteristics that the surface chemical characteristic algorithms 120 can predict can include an oxidation rate, a reactivity rate, a chemical composition ratio, or a moisture evaporation rate.

[0054] In particular, the surface chemical characteristic algorithms 120 can determine the surface chemical characteristics of the particles based on the measurement data 132. The surface chemical characteristic algorithms 120 can identify regions of pixels of the images corresponding to the particles, and accordingly, can identify one or more chemicals on the surface of the particles.

[0055] In particular, the controller 106 can generate feature vectors of the pixel regions as input, and the controller 106 can provide the feature vectors to the surface chemical characteristic algorithms, specifically, to a machine learning model of the algorithms 120 to identify the chemicals. Based on identifying the chemicals, the controller 106 can generate a chemical composition ratio for the particles. The chemical composition is a percentage content of each of the identified chemicals.

[0056] In some examples, based on the particles 130 being wetted particles, the controller 106 can determine a moisture evaporation rate at the surface of the wetted particles. In particular, the sensor rig 116 can provide measurement data 132 including images of the wetted particles from an infrared sensor (e.g., a NIR camera), where the images track the surface moisture of the wetted particles over time based on multiple sensors. In some examples, the measurement data 132 can include infrared images of the particles that map a temperature profile of the surface of the wetted particles to determine the moisture evaporation rate. In some examples, the system can determine a porosity of the surface of the wetted particles based on the measurement data 132.

[0057] The physical characteristic algorithms 122 can determine physical characteristics of particles given measurement data 132. For example, a machine learning model of the physical characteristics algorithms 212 can be trained to determine physical characteristics of particles given multiple types of measurement data. Some characteristics that the machine learning model can predict can include particle size, particle shape, particle volume, surface area, particle sphericity’, elemental composition, mineralogy, surface texture, and mechanical properties.

[0058] The surface chemical characteristic algorithms 120 and the physical characteristic algorithms 122 can be trained by the controller 106. For example, the controller 106 can use the algorithms 212 before using a machine learning model to determine characteristics of particles. In particular, each machine learning model can obtain and process training data. For example, the training data can be labeled training data that includes sets of measurement data from multiple sensors of the same batch of particles. Each set of measurement data can correspond to a different batch of particles, for example.

[0059] For example, each label can indicate a characteristic such as particle size, particle shape, particle volume, surface area, particle sphericity, elemental composition, mineralogy, surface texture, and mechanical properties of particles in the batch of particles. In some examples, the labels can indicate the average or majority characteristic of particles in the batch of particles. For example, where a batch of particles includes particles with different elemental compositions, the label for the characteristic of elemental composition can be the elemental composition that is most common for the particles in the batch.

[0060] In some implementations, the machine learning models can include one or more models. For example, the physical characteristic algorithms 122 can include a segmentation model that has been trained to extract information about individual particles from images of the measurement data 132. The algorithms 212 can also include a vision transformer (ViT) that has been trained to determine information about geometry of the individual particles from the information extracted by the segmentation model. In some implementations, the algorithms 212 can also include a regression layer to determine characteristics of the particles from information determined by the ViT. In some implementations, the algorithms 212 can also include a softmax layer for classifying rock gradations.

[0061] In some implementations, the algorithms 212 determine characteristics of particles based on the type of particle given measurement data 132. In particular, a machine learning model of the algorithms 212 can be trained to correlate characteristics with particle type and measurement data from particular types of sensors, such as the sensors described in furtherdetail with reference to FIG. 2. For example, the controller 106 can determine the type of particle and provide the type of particle as another input to the machine learning model of the algorithms 212. In some implementations, the controller 106 can determine the type of particle using another machine learning model. In some implementations, the controller 106 can receive an input from the user defining the ty pe of particle. In some implementations, the controller 106 can use a different machine learning model for each different type of particle. For example, each machine learning model can be trained on training examples of only a specific type of particle.

[0062] The controller 106 is configured to provide the particles 130 to an example particle processing systems 108. In particular, the controller 106 can control the conveyor belt 104 to provide the particles 130 to the example particle processing systems 108. In some examples, the controller 106 can control a set of conveyor motors to operate the conveyor belt 104. In some examples, the controller 106 can adjust a speed of the conveyor belt to improve sensor measurement of measurement data 132.

[0063] The controller 106 can output characteristics 134 corresponding to the provided particles 130 to particular components, such as the example particle processing systems 108. The characteristics 134 can include the surface chemical characteristics generated by the surface chemical characteristics machine learning model 120, the physical characteristics generated by the physical characteristic machine learning model 122, or both. The example particle processing systems 108 can be a carbonation systeml24, a densification system 126, and a blast furnace system 128, as described in further detail below with reference to FIGs. 3A and 3B.

[0064] The controller 106 can output different characteristics to relevant different components along the conveyor belt. In order to output characteristics 134 relevant to each component, the controller 106 can determine which of the characteristics outputted from the machine learning model of the algorithms 120 or 122 is relevant to each of the example particle processing systems 108. In particular, the controller 106 can determine the relevance of the characteristics based on one or more lookup tables, as described in more detail below with reference to FIG. 4.

[0065] In some examples, the controller 106 can send control signals 136 to the example particle processing systems 108. Operations of the example particle processing systems 108 can be controlled by the control signals 136. In particular, control signals 136 from the controller 106 can control operations of the carbonation system 124, the densification systeml26, and the blast furnace

[0066] For example, the controller 106 can provide the characteristics to the blast furnace systeml28. The blast furnace system 128 is configured to process the particles 130 to generate smelted iron using the controller 106 based on the characteristics 134, as described in further detail below with reference to FIG. 3A.

[0067] In another example, in order to generate a concrete mixture, the controller 106 can provide the characteristics to the carbonation system 124, the densification systeml26, or both. The carbonation subsystem 124 and the densification subsystem 126 are configured to process the particles 130 to generate a concrete mixture using the controller 106 based on the characteristics 134, as described in further detail below with reference to FIGs. 3 A and 3B.

[0068] FIG. 2 is a top view of an example system 200 for determining surface chemical characteristics of particles on a conveyor belt 104. The particles can be, for example, rocks. The system 200 can include the conveyor belt 104, a first section of the conveyor belt, a sensor rig 116, a controller 106, a lighting system 114, and a sprayer system 118.

[0069] The lighting system 114 includes multiple lighting devices arranged to direct light at a specific angle with respect to at least one of the sensors. In some examples, the lighting system 114 includes IR and / or ultraviolet (UV) illuminators to direct IR / UV light to the particles. For example, the IR / UV light can excite different materials within the particles to emit light at particular frequencies and intensities that are characteristic of those materials. The IR sensors (e.g., the NIR sensors or the SWIR sensors) can detect the emissions and identity’ the materials based on the emitted light frequencies and intensities.

[0070] The sensor rig 1 16 can be located proximate to the first section. The sensor rig 116 can include different types of sensors arranged to obtain measurements of materials conveyed by the first section of the conveyor belt 104. The different types of sensors can be arranged in different locations within the sensor rig 116.

[0071] The sensor rig 116 includes one or more sensors 202 configured to capture the measurement data 132. The sensors 202 are arranged on the sensor rig 116 to obtain measurements of particles conveyed on the first section. For example, the sensor rig 116 can extend along the first section. Sensors 202 can be arranged on, over, around, or to the side of the first section. For example, sensors 202A and 202B are located on the side of the first section, and sensor 202c is positioned over the first section. In the example of FIG. 2, particles on the first section are conveyed in the direction of sensor 202A to sensor 202C.

[0072] The first section can also include mechanisms arranged to create environmental conditions favorable to sensor reading. Environmental conditions favorable to sensor reading can include favorable or ideal lighting or illumination conditions or temperature conditions.Lighting or illumination conditions can be affected by brightness, color temperature, or wavelength of light, for example. An example of a mechanism to create lighting conditions favorable to sensor reading is a lighting device 206 arranged to direct light at a specific angle with respect to at least one of the sensors 202. For example, the lighting device 206A can be arranged to direct light at a specific angle with respect to sensor 202a. For example, if sensor 202A is a visible light camera, lighting device 206A can be a visible lighting device. The lighting device 206A can be arranged to direct visible light at the first section at the region that sensor 202A is directed to.

[0073] The lighting device 206 can also have a particular brightness or color temperature. The lighting device 206A can thus provide ideal illumination conditions for the sensor 202A In some implementations, the lighting device 302a can be arranged to direct visible light at the environment of the first section. For example, the lighting device 302a can illuminate the environment to provide certain illumination conditions for the sensor 202a.

[0074] As another example, if sensor 202b is an IR camera, lighting device 302b can be an IR lighting device. The lighting device 302b can be arranged to direct IR light at the first section at the region that sensor 202b is directed to. The lighting device 302b can thus provide ideal illumination conditions for the sensor 202b.

[0075] As another example, if sensor 202b is a telecentric camera sensor, lighting device 302b can be a device that creates conditions of sharp contrast, such as a backlight. The lighting device 302b can be arranged to direct bright light at the first section at the region that sensor 202b is directed to. The lighting device 302b can thus provide ideal illumination conditions for the sensor 202b.

[0076] As another example, if sensor 202b is a stereo camera sensor, lighting device 302b can be arranged to direct light from the direction of the sensor 202b to create illumination conditions for disparity algorithms for depth sensing. The lighting device 302b can be arranged to direct light at the first section at the region that sensor 202b is directed to, from the direction of the sensor 202b. The lighting device 302b can thus provide ideal illumination conditions for the sensor 202b.

[0077] In some implementations, the controller 106 can adjust the settings of the lighting devices 206. For example, the controller 106 can adjust the brightness, warmth, wavelength, and angle of each lighting device 206. For example, the controller 106 can receive data indicating the illumination conditions for the images that the machine learning model 212 was trained on. The controller 106 can adjust the settings of the lighting devices 206 to more closely match the illumination conditions.

[0078] In some implementations, the sensor rig 116 can include multiple sensors 202 of the same type, and at least some of the sensors 202 can correspond to lighting devices 206 of different settings. For example, the multiple sensors 202 of the same type can be visible light cameras. Each lighting device 206 can be arranged to direct light at a different angle with respect to the visible light cameras. For example, the lighting devices 206 can be arranged to cast different types of shadows, e.g.. of different intensity or in different directions. The machine learning model 212 can infer information that can affect its predictions from regions of the images that include different shadows.

[0079] In some implementations, the first section can include lighting devices and background features that create a high contrast background for sensor measurements. For example, background features can include the color of the first section. In some examples, the particles can have a lighter color. To create a high contrast background, the color of the first section can be a dark color and the lighting devices can have a high brightness. In some examples, the particles can have a darker color. To create a high contrast background, the color of the first section can be a light color and the lighting devices can have a high brightness. The high contrast background can make it easier for the machine learning model 212 to separate regions of an image that include a particle from regions of an image that include the first section, for example. The high contrast background can also make it easier for the machine learning model 212 to separate regions of an image that include a particle from regions of an image that include a shadow of the particle.

[0080] In some examples, the conveyor belt 104 includes a sprayer system 1 18. The sprayer system can be located proximate to the first section. The sensor rig 116 can include different types of sprayers 204 arranged to spray materials conveyed by the first section of the conveyor belt 104 with one or more liquids, chemicals, or both. The different types of sprayers 204 can be arranged in different locations within the sensor rig 116.

[0081] The sensor rig 116 can send the measurement data 132 to the controller 106 at a regular interval. For example, the controller 106 can cause particles to enter the first section at regular intervals of time. Sensor 202a can thus obtain measurement data 132 when particles are being conveyed past sensor 202a at the regular intervals of time. Obtaining measurement data 132 and determining characteristics at a regular interval can provide for real-time monitoring that allows the system to react to changes in the particles quickly and efficiently. In addition, downstream decisions such as adjusting a region for mining, mixture recipe, blast furnace reactants, or crusher setting can be made more quickly and efficiently in response to changes in the particles.

[0082] The controller 106 can obtain the measurement data 132 from the conveyor belt 104. In some implementations, the controller 106 can also obtain metadata from the sensors 202. For example, the metadata can include information such as a timestamp of when each measurement in measurement data 132 was taken, or a sensor identifier that identifies which sensor of the sensors 202 took each measurement in measurement data 132. The controller106 is configured to analyze the measurement data and communicate with components, such as the example particle processing systems 108. on the conveyor belt 104. For example, the controller 106 can send instructions to the components, such as the example particle processing systems 108, based on analyses of measurement data obtained from the sensors. For example, the controller 106 can apply the measurement data to a machine learning model configured to determine characteristics of particles given measurement data as described below with reference to FIGs. 3A and 3B.

[0083] In some implementations, the controller 106 can send a request to each of the sensors 202 at different times. For example, in implementations where the sensors 202 are arranged along the first section, as depicted in FIG. 2. different sensors may be directed at a particular batch of particles at different times. The controller 106 can determine estimated times for each sensor that the particular batch of particles will be located at the region of the first section that the sensor is directed at.

[0084] For example, the controller 106 can use sensor metadata, sensor position data, and conveyor operation data to determine an amount of time it would take for the batch of particles to travel to each sensor. The controller 106 can send a request to each of the sensors 202 to obtain measurement data at a specified time based on the amount of time for each sensor. For example, if it takes one minute for the batch of particles to travel from sensor 202a to sensor 202b. and one minute for the batch of particles to travel from sensor 202b to 202c, the system can send a request to sensor 202a to take a measurement at 12:00, a request to sensor 202b to take a measurement at 12:01, and a request to sensor 202c to take a measurement at 12:02. In some implementations, the system can send multiple requests to each sensor. For example, the system can send a request for measurement data at regular intervals of time. For example, the system can send a request to sensor 202a to take measurements at 12:00 and 12: 10, a request to sensor 202b to take measurements at 12:01 and 12:11, and a request to sensor 202c to take measurements at 12:02 and 12: 12.

[0085] The system 100 can determine characteristics of particles or materials being conveyed on the conveyor belt 104. The materials may be included in a material mixture or final product. For example, the materials can be rocks or particulate materials to be includedin a concrete mixture. The system 100 can be part of a material preparation system. For example, a material preparation system can use the system 100 to adaptively adjust proportions of materials in the material mixture. The system 100 can adjust the proportions based on characteristics determined by the system 100 to achieve a set of target properties of the mixture, for example. Target properties of the mixture can include, for example, rheological properties (e g., flow characteristics) of the mixture. Target properties of the final product can include, for example, structural properties, electrical properties, thermal properties, chemical properties, or the like, of a final product. In some examples, diffuse reflectance spectroscopy can be used across the visible, near- and shortwave-infrared spectral regions (400 to 2500 nm) as a tool to assess the strength of particles.

[0086] Ilmportantly, the sensors 202 are configured to capture the measurement data 132, and the system can provide the measurement data 132 to the controller (e.g., to the larger system) to determine the surface chemical characteristics of particles. The surface chemical characteristics can include, but are not limited to, an oxidation rate, a reactivity rate, a chemical composition ratio, or a moisture evaporation rate.

[0087] Additionally, the sensors 202 are configured to capture the measurement data 132, and the system can provide the measurement data 132 to the controller to determine physical characteristics of particles. Importantly, in some examples, the system can use the physical characteristics to determine the surface chemical characteristics of the particles with increased accuracy, as described in further detail below with reference to FIG. 5. The physical characteristics of particles can include particle size, particle shape, particle volume, surface area, and particle sphericity. Characteristics of particles can also include elemental composition, mineralogy, surface texture, and mechanical properties, such as water content, specific surface area, or surface roughness.

[0088] In some implementations, the system can determine characteristics for one particle at a time. In some implementations, the system can determine characteristics for a batch of particles. For example, the system can determine characteristics that represent a majority of the particles in the batch, or an average of the particles in the batch. For example, particles in a batch may have different surface areas. The system can determine a surface area for the batch based on the average surface area of the particles in the batch. As another example, particles in the batch may have different elemental compositions.

[0089] The system can determine an elemental composition for the batch based on the elemental composition of the majority of the particles in the batch, for example. In some examples, the sensor rig 116 is configured to send measurement data 132 of the top surface ofthe wetted particle. Based on the measurement data, the system can determine a moisture evaporation rate at the surface of the wetted particles, or a batch of the wetted particles.

[0090] The conveyor belt 104 can be configured to convey materials located on the conveyor belt. The conveyor belt 104 can be part of a larger material preparation system and can convey materials to and from other points in the material preparation system, or to an example particle processing systems 108.

[0091] The controller 106 can control the mechanism to cause particles on the conveyor belt 104 to enter the first section at intervals of time by adjusting a speed of the conveyor motors of the conveyor belt 104.

[0092] Importantly, controller 106 can output relevant characteristics to each component. For example, the controller 106 can determine that a characteristic is relevant for the operations of a particular component, such as a blast furnace system 128, but not for the operations of a different component, such as a carbonation system 124. The controller 106 can send the characteristic to the particular component, as described in further detail below with reference to FIGs. 3A and 3B.

[0093] The controller 106 can communicate with the components and the sensors using a communications interface. The communications interface is a network interface, e.g., a cellular network interface, a local area network interface (e.g., Wi-Fi interface), a fiber optic interface, or another appropriate networking interface.

[0094] FIGs. 3A and 3B show exemplary particle processing systems for processing particles, e.g., minerals, crushed concrete for recycling, ore pellets for refining, etc, that incorporate the system 100 (described above) for determining surface chemical characteristics. FIG. 3A depicts an example of the concrete generation system, which includes a carbonation system 124 and a densifi cation system 126.

[0095] In some examples, the system processes materials 302 in a physical processing system 110, such as the crusher 304 or a pelletizing system. The crusher 304 can crush the materials 302 to particular sizes and / or geometries. Operations of the crusher 304 can be controlled by control signals 136 from the controller 106. For example, control signals 136 from the controller 106 can cause the crusher to increase or decrease the size of particles 130.

[0096] Importantly, the conveyor belt 104 includes sensors 202 from above that are configured to extract measurement data 132 from the particles in order to determine characteristics of the particles 130, such as surface chemical characteristics, physical characteristics, or both. That is, the system is configured to provide the characteristics to theparticle processing system, and the system can use the data to inform its own operations (e.g., operating the blast furnace).

[0097] The sensors 202 are arranged proximate to the conveyor belt 104 to obtain measurement data 132 of the particles 130. For example, in some implementations optical sensors can be arranged in an array along the conveyor belt 104 or a chute used to convey the particles to the example particle processing systems 108. The optical sensors can transmit images of the particles to the controller 106. which (as explained in more detail below) can use image processing algorithms to identify particle surface chemical characteristics, particle physical characteristics, or both.

[0098] The controller 106 uses the measurement data 132 to determine characteristics of the particles 130. For example, particle characteristics can include, but are not limited to, surface chemical characteristics of particles which include, but are not limited to, an oxidation rate, a reactivity rate, a chemical composition ratio, or a moisture evaporation rate. The physical characteristics of particles can include particle size, particle shape, particle volume, surface area, and particle sphericity. Characteristics of particles can also include elemental composition, mineralogy, surface texture, and mechanical properties, such as water content, specific surface area, or surface roughness.

[0099] In some examples, rheometry measurements are performed on an initial concrete mixture made from the upgraded particles 130. Rheometry measurements of the concrete mixture can be estimated based on the measured characteristics of the particles. The rheometry measurements are used to predict characteristics of the concrete after curing. The actual rheometry7measurements of the concrete mixture can be obtained and compared with the estimated rheometry to determine whether to adjust an amount of additives 312. The system can determine, based on the rheometry measurements, whether the concrete mixture is likely to achieve a desired set of post-curing characteristics. If not, the initial mixture is adjusted through an iterative process until the rheometry measurements indicate that the concrete mixture is likely to achieve the desired post-curing characteristics.

[0100] The controller 106 can send control signals 136 to control an amount of CO2 and H2O provided to the particles 105 by the carbonation system 115. The controller 106 can also send control signals 136 to control an amount of additives 312 provided to the particles 105 by the densifi cation system 126. By adaptively maximizing the degree to which additive reactions accrue, carbon uptake by the particles can be increased. Additionally, compression strength of heterogeneous particle mixtures can be enhanced.

[0101] In some examples, the sensors 202 can analyze the output treated aggregate, e.g., upgraded particles 130, and can provide feedback to the controller 106. Based on the feedback from the sensors 202, the controller 106 can use control signals 136 to adjust one or more of a size of particles 105 crushed by the crusher, an amount of CO2 106, an amount of H20 108, or an amount of additives 312 to improve the characteristics of the upgraded particles 130. The upgraded particles 130 can be mixed into a concrete mixture. Concrete mix sensors provide rheometry measurements of the concrete mixture to the controller 106. For example, the concrete mix sensors can measure various attributes of the concrete mixture that can be used to estimate or compute rheumatic properties of the concrete mixture in real-time. The concrete mix sensors can include, but are not limited to, viscosity' sensors, rheometers, temperature sensors, moisture sensors, ultrasonic sensors (e.g., ultrasonic pulse velocity sensors), electrical property sensors (e.g., electrodes, electrical resistance probes), electromagnetic sensors (e.g., short-pulse radar), or other sensors (e.g., geophone, accelerometer). The concrete mix sensors can include, but are not limited to, hydrophobicity, moisture content, XRD spectra, XRF spectra, static yield stress, acoustic impedance, p-wave speed, dynamic yield stress, static modulus of elasticity, Young’s modulus, bulk modulus, shear modulus, dynamic modulus of elasticity (DME), Poisson’s ratio, density, resonance frequency, nuclear magnetic resonance (NMR), dielectric constant, electric resistivity', polarization potential, and capacitance.

[0102] In some implementations, the system may include a metering hopper. The metering hopper may be used to collect and measure (e.g., weigh) particles as they pass through the sensors 202. For example, the weight of the particles measured by a metering hopper can be passed to the material processing system 100 permitting the control system to monitor the weight of the particles being measured in real-time. In some implementations, material processing system 100 can be retro-fit to a traditional ready-mix concrete plant. For example, adding the material processing system 100 to a ready -mix plant may allow the ready-mix plant to more precisely tailor concrete mixes for specific applications and job sites.

[0103] Particles can be conveyed from the crusher to the carbonation system 115 and to the densification system 126. For example, the particles 105 can be conveyed by a series of conveyors and augers. The particles 105 are passed through the sensors 202 prior to delivery' to the blast furnace system 128, prior to delivery' to carbonation system 115, in between the carbonation system 115, and after departing from the densification system 126.

[0104] Additionally or alternatively, the concrete generation system includes the carbonation system 124, the densification system 126, or both. The controller 106 can controlthe operations of one or more ingredient metering systems based on analyses of data obtained from the sensors 202, such as CO2 308, H20 310, and additives 312.

[0105] The carbonation system 124 performs a process for accelerated carbonation of the particles 130. Based on the characteristics of the particles 130, the carbonation system 115 determines suitable process conditions for the accelerated carbonation. The characteristics can include surface chemical characteristics, such as moisture evaporation rate, calcium hydroxide content , or moisture evaporation rate, the physical characteristics, such as size, surface area, shape, and porosity, or both. The process conditions include, for example, temperature, water vapor 310, and CO2 concentration 308.

[0106] A purpose of the accelerated carbonation process is to store the largest possible amount of CO2 in the particles in order to improve particle properties. Carbonation decreases the water absorption coefficient by filling pores due to the formation of calcium carbonates. Thus, carbonation leads to the formation of calcium carbonates and to a decrease in total porosity. The capillary porosity is decreased due to clogging of the pores. In addition, carbonation increases the microporosity of particles as a result of decalcification and mercury intrusion.

[0107] The densification system 126 performs a densification process. Based on characteristics of the particles after carbonation, the densification system 126 can apply suitable amounts and types of additives 312. The additives 312 can include, for example, silicate sources and catalysts to maximize reactivity. The densification process can improve the quality of particles by using pozzolanic and sodium silicate solution as treatment solutions. Sodium silicate combined with pozzolanic materials can improve mechanical properties of particles. For example, a solution of sodium silicate and silica fumes can improve compressive strength of particles.

[0108] The densification system 126 produces upgraded particles 130. Upgraded particles 130 can be particles that have undergone a carbonation process, a densification process, or both. A post processing characterization stage can be performed using the same reactivity7estimation and other optically determined physical characteristics to provide an accurate qualification of the upgraded particles' compression strength, porosity, uniformity, and other physical characteristics. This measure can allow for quality control by providing insights into material strength, water absorption, and flowability.

[0109] For example, viscosity, moisture, and temperature sensors can be used to measure rheologic properties of the concrete mixture such as changes in the viscosity of the mixture over time and at different moisture content levels and temperatures. As described in moredetail below, the controller 106 can use the rheometry measurements to determine whether and how much additional ingredients and / or additives should be added to the concrete mixture to obtain desired concrete properties.

[0110] During the iterative adjustment process, upgraded particles 130 are incrementally added to the initial concrete mixture while changes in the rheometry' measurements are monitored. Additional upgraded particles 130 are added until the rheometry measurements indicate that the concrete mixture is likely to achieve the desired post-curing characteristics. Such post-curing characteristics can include, but are not limited to, compressive strength, tensile / flexural strength, flowability, toughness, cure time, cure profile, finish, density (wet & dry), thermal insulation, shrinkage, and slump.

[0111] Post-curing characteristics can be determined from rheometry measurements by, e.g., using multi-dimensional lookup tables relating experimentally obtained post-curing characteristics to mixtures with known rheological properties, by applying theoretical and analytical particle packing model-based Bay esian optimization algorithms to the rheometry measurements, or a combination thereof.

[0112] In some examples, the post-curing characteristics can be provided as feedback to the controller 106. Based on the feedback, the controller 106 can use control signals 136 to adjust one or more of a size of particles 130 crushed by the crusher 304, an amount of CO2 308, an amount of H20 310, or an amount of additives 312 to improve the characteristics of the cured concrete mixture.

[0113] FIG. 3B depicts an example blast furnace system. In operation, the system processes materials 302 to generate the particles 130, and system uses the controller 106 to provide the particles 130 to the blast furnace system 128 using the conveyor belt 104, where the conveyor belt includes one or more sensors 202 that capture measurement data 132 and send the measurement data to the controller 106 to generate control signals 136. The system can achieve desired properties of the particles 130 based on the measurement data to provide the particles to the blast furnace system 128[H4]

[0115] In some examples, the system processes materials 302 in a physical processing system 110, such as the crusher 304 or a pelletizing system. The crusher 304 can crush the materials 302 to particular sizes and / or geometries. Operations of the crusher 304 can be controlled by control signals 136 from the controller 106. For example, control signals 136 from the controller 106 can cause the crusher to increase or decrease the size of particles 130.The conveyor belt 104 includes sensors 202 that measure characteristics of the particles 130, such as surface chemical characteristics, physical characteristics, or both.

[0116] Importantly, the conveyor belt 104 includes sensors 202 from above that are configured to extract measurement data 132 from the particles in order to determine characteristics of the particles 130, such as surface chemical characteristics, physical characteristics, or both. That is, the system is configured to provide the characteristics to the particle processing system, and the system can use the data to inform its own operations (e.g., processing the concrete).

[0117] The sensors 202 are arranged proximate to the conveyor belt 104 to obtain measurement data 132 of the particles 130. For example, in some implementations optical sensors can be arranged in an array along the conveyor belt 104 or a chute used to convey the particles to the example particle processing systems 108. The optical sensors can transmit images of the particles to the controller 106, which (as explained in more detail below) can use image processing algorithms to identify particle surface chemical characteristics, particle physical characteristics, or both.

[0118] The controller 106 uses the measurement data 132 to determine characteristics of the particles 130. For example, particle characteristics can include, but are not limited to, surface chemical characteristics of particles which include, but are not limited to, an oxidation rate, a reactivity rate, a chemical composition ratio, or a moisture evaporation rate. The physical characteristics of particles can include particle size, particle shape, particle volume, surface area, and particle sphericity. Characteristics of particles can also include elemental composition, mineralogy7, surface texture, and mechanical properties, such as water content, specific surface area, or surface roughness.

[0119] In some examples, such as for iron ore smelting using the blast furnace system 128, measurements are performed on the particles 130 to determine whether to adjust reactants 306 or control aspects of the blast furnace, as described in further detail below.

[0120] The particle analysis can be used to optimize the additive reactant and process parameters. For example, the controller 106 can send control signals 136 to control an amount of reactants 306 provided to the particles 130 by the blast furnace system 128. Particles can be conveyed from the crusher to the blast furnace system 128. For example, the particles 105 can be conveyed by a series of conveyors and augers. The particles 105 are passed through the sensors 202 prior to delivery to the blast furnace system 128. In some examples, the particles can pass through only the blast furnace system 128.

[0121] The blast furnace system 128 can generate smelted iron by processing the particles 130. Operations of the blast furnace system 128 can be controlled by the controller 106. The controller 106 receives measurement data 132 from the sensors 202. The controller 106 can control the operations of one or more metering systems based on analyses of data obtained from the sensors 202. In some examples, the controller 106 can be used to control the operations of one or more metering systems for other liquid solid or solid gas-reactions like catalytic reactors, or charges for electric arc furnaces and oxygen furnaces.

[0122] The blast furnace system 128 includes a first section and a second section of a blast furnace. The inputs of the blast furnace system 128 enter from the top of the blast furnace. Generally, the inputs include iron ore pelletized particles (e.g., particles 130) and reactants 306, such as coke and limestone.

[0123] In particular, the blast furnace system 128 passes the input to the first section to generate a reduction of iron ore at high temperatures (e.g., 250°C - 750°C). The blast furnace system 128 then passes the products of the reduction reaction to the second section, which is at relatively higher temperatures (750°C - 1500°C) than the first section. In the second section, the inputs reacts with CO2. In particular, the coke reacts with the CO2. Additionally, the limestone decomposes and produces slag. The blast furnace system 128 then adds blast air at a certain flow rate and blast air temperature, which reacts with the coke to expel waste gases. Thus, at the bottom of the blast furnace, smelted molted iron remains as an output of the blast furnace system 128.

[0124] The surface chemical characteristics of the particles 130 can have a direct effect on the surface reaction rates at the second section of the blast furnace system, as variability in the composition of the chemicals of the particles can increase or decrease the surface reaction rates. In order to optimize the reactivity of the particles 130, the controller 106 can use the measurement data 132 to determine a chemical composition ratio for the particles. In some examples, the controller 106 can further determine an oxidation rate, a reactivity rate, or a moisture evaporation rate. Based on these surface chemical characteristics, the controller 106 can adaptively adjust the proportion of reactants 306, adjust a flow rate of the blast air, or adjust a temperature of the blast air added to the particles 130 based on the measured characteristics.

[0125] Additionally, the surface reaction rate of the blast furnace system 128 depends on the mixability' of the different reactants 306 and the particles 130, especially the particles 130. In particular, the homogeneity of the particles has a direct relationship with the surface reaction rates at the second section of the blast furnace system 128. The homogeneity of theparticles depends on the similarity of particle size for each of the particles. For example, particles of the same size tend to aggregate, which can lead to higher surface reaction rates, and particles of different sizes tend to segregate, which can lead to inefficient reactions and relatively lower surface reaction rates.

[0126] On the other hand, the homogeneity7of the particles has an inverse relationship with the packing density of the particles. Particles of similar sizes can create voids in the flow of particles 130 that can be too large, resulting in under packing, or too small, which can decrease blast furnace efficiency and result in incomplete combustion of the coke. Thus, there is a tradeoff between mixability and packing density based on the homogeneity of the particles. Higher packing density can be achieved by exponential size distribution (heterogeneity), whereas higher mixability is achieved by greater size homogeneity7.

[0127] Additionally, the shape of the particles 130 can affect the efficiency of the blast furnace system 128. Different particle shapes can also create voids in the flow of particles 130 that are too large or too small, which can affect the way that gas (e.g., CO2 ) flows through the blast furnace system 128, which impacts the quality of the smelted molten iron. The particle shape also affects the surface area of the particles, which has a direct impact on the surface reaction rates of the blast furnace system 128 and the way the particles 130 move through the blast furnace system 128.

[0128] In order to optimize the mixability and the packing efficiency of the particles 130, the controller 106 can use the measurement data 132 to calculate a shape-irregularity factor (STF) based on measurements from images of each particle, including physical characteristics, such as size, shape, volume, and specific surface area. For each particle, the SIF represents a specific excess surface area of an irregular shape of the particle compared to regular spheres of the same size or volume as the particle.

[0129] For example, a particle within a certain size range (e.g., 2mm-3mm) can have an irregular shape. The sensors 202 can capture images of the particle and provide the images to the controller 106 as part of the measurement data 132. The controller 106 can measure a size, shape, and specific surface area of the particle, and the controller 106 can determine a SIF of the particle based on comparing the irregularities of the particles with other particles within the same size range.

[0130] As such, the controller 106 can analyze the particles 130 using measurement data 132 from the sensors 202 by determining a SIF for each of the particles 130, as described in further detail with reference to FIG. 5. The particle analysis can be used to optimize the iron ore smelting process. For example, the controller 106 can send control signals 136 to controlan amount of reactants 306 and particles 130 added to the blast furnace system 128, and the controller 106 can send control signals 136 to control parameters of the blast air, such as a flow rate of the blast air, a temperature of the blast air, or both.

[0131] In some examples, the controller 106 can use control signals 136 to adjust control parameters of the physical processing system 110. For example, the control parameters include one or more of a size of particles 130 generated by the pelletizing drum to improve the characteristics of the particles 130 based on the SIF of each particle. In particular, the controller 106 can determine, for a certain number of particles, that the SIF is outside of a specified range, and the controller 106 can send one or more control signals 136 to the pelletizing drum.

[0132]

[0133] FIG. 4 is a block diagram of an exemplary controller 106 for the material processing system 100. The controller 106 includes a computing system 402 in communication with blast furnace control 406, concrete mixing control 408, particle analysis sensors 202, and a metering control system 208 which can control the control signals 136 for additives and reactants. Computing system 402 is configured to control various aspects of the blast furnace iron ore smelting process, the recycled concrete preparation process, or both. For example, computing system 402 can store and execute one or more computer instruction sets to control the execution of aspects of the blast furnace iron ore smelting process, the recycled concrete preparation process, or both described herein. Computing system 402 can include a system of one or more computing devices. The computing devices can be, e g., a system of one or more servers. For example, a first server can be configured to receive and process data from the concrete mix sensors and the particle analysis sensors 202. Another server can be configured to interface with the metering control system 410 and issue control commands based on analysis results from the first server.

[0134] In some implementations, the computing system 402 can be operated or controlled from a user computing device 404. User computing device 404 can be a computing device, e.g., desktop computer, laptop computer, tablet computer, or other portable or stationary computing device.

[0135] Briefly, computing system 402 can control the material processing system 100 to prepare materials for the blast furnace iron ore smelting process, the recycled concrete preparation process, or both. In some examples, the computing system 402 controls the blast furnace subsystem for iron ore smelting. In some other examples, the computing system 402 controls the concrete generation system including the carbonation subsystem and thedensification subsystem for generating prepared concrete mixtures. The computing system 402 can use the particle analysis sensors 202 to determine surface chemical characteristics of particles. In some examples, the computing system can use the particle analysis sensors 202 to determine physical characteristics of particles.

[0136] In some implementations, control system 402 can control a combined system that iteratively adjusts control parameters of the pre-processing subsystem based on the characteristics from measurement data 132 of the sensors 202. For example, the control system 402 can control parameters of a pelletizing drum of the particles 130 from images and measurements captured at the output of the pelletizing drum.

[0137] The computing system 402 can send control signals 136 to the pelletizing drum based on the determined adjustments to the control parameters. For example, the computing system 402 can issue commands to the metering control system 410 to control the addition of additives to the particles 130, such as CO2 and / or H2O. In particular, the computing system 402 can then interface with the metering control system 410 to operate valves from clay and / or H2O supply tanks of additives in order to apply appropriate amounts of clay and / or H2O to the pelletizing drum.

[0138] In some other examples the computing system 402 can interface with the pelletizing drum to adjust the crush size of the particles, adjust the spin speed of the pelletizing drum, the feed rate of the pelletizing drum, or a combination thereof.

[0139] In some examples, the computing system 402 can send control signals 136 to the conveyor belt 104 to convey the particles 130 to an example output system.

[0140] The computing system 402 can then send control signals 136 to control the blast furnace subsystem via the blast furnace control 406. to control the concrete generation system via the mixing control 408, or both.

[0141] In some implementations, computing system 402 can include a set of operations modules 412 for controlling different aspects of the example particle processing systems 108. The operation modules 412 can be provided as one or more computer executable software modules, hardware modules, or a combination thereof. For example, one or more of the operation modules 412 can be implemented as blocks of software code with instructions that cause one or more processors of the computing system 402 to execute operations described herein. In addition or alternatively, one or more of the operations modules can be implemented in electronic circuitry' such as, e.g., programmable logic circuits, field programmable logic arrays (FPGA). or application specific integrated circuits (ASIC). Theoperation modules 412 can include an additive and reactant controller 414, a particle analyzer controller 416, estimation algorithms 418. and one or more lookup tables 420.

[0142] Additive and reactant controller 414 interfaces with the metering control system 410 to control the addition of reactants to the blast furnace subsystem 128 and additives to the carbonation subsystem 124 and densification subsystem 126. For example, the additive and reactant controller 414 can issue commands from the computing system 402 to the metering control system 410 to control the addition of additives to the particles 130 in the carbonation subsystem 124, the densification subsystem 126, or both.

[0143] Particle analyzer control 416 interfaces with the particle analysis sensors of the sensors 202. Particle analyzer controller 416 receives and buffers data from the particle analysis sensors 202. The particle analyzer controller 416 can process the sensor data to determine particle characteristics of each analyzed particle. For example, as discussed in more detail below, the particle analyzer controller 416 can execute data analysis algorithms to interpret the sensor data and determine particle characteristics including, but not limited to, surface chemical characteristics, particle size distributions, particle shape distributions, and particle surface area distributions.

[0144] The control system can employ estimation algorithms to estimate the characteristics of the particles. In some examples, the estimation algorithms 418 include algorithms to estimate surface chemical characteristics of the particles. In particular, the estimation algorithms 418 can determine the surface chemical characteristics of the particles by identifying regions of pixels of the images from the measurement data 132. The system can identify one or more chemicals on the surface of the particles based on identify ing the regions of the pixels, and the system can determine a chemical composition by employing the estimation algorithms 418. In some examples, such as analysis of wetted particles, the control system can employ estimation algorithms 418 to determine a moisture evaporation rate of the particles based on the identified pixels of the measurement data. In some other examples, the control system can employ the estimation algorithms to determine one or more of an oxidation rate or a reactivity rate based on the identified pixels.

[0145] In some examples, the control system can employ the estimation algorithms 418 to estimate rheometry parameters based on the particle characteristics of the particles. For example, the estimation algorithms 418 can employ lookup tables 420 to determine estimated rheometry measurements. The computing system can include a lookup table 420 that correlates concrete particle characteristics to experimentally determined rheometry parameters. In some implementations, the rheometry estimation algorithms includealgorithms that estimate particle packing efficiencies from the particle parameters and a lookup table 420 that correlates particle packing efficiencies with experimentally determined rheometry parameters. The computing system 402 can then compare the estimated particle packing efficiencies to the data in the lookup table 420 to estimate the rheometry parameters of particles. In some implementations, rheometry7estimation algorithms 418 include a packing efficiency model to determine a packing efficiency of the mixture based on the particle characteristics. The model can be a theoretical and analytical particle packing modelbased Bayesian optimization algorithm-or other machine learning model — to determine a packing efficiency of the particles and estimate rheometry7parameters of the mixture.

[0146] In some implementations, the estimation algorithms 418 can include a machine learning model to estimate surface chemical characteristics (e.g., surface chemical characteristic machine learning model 120), such as chemical composition ratio, or physical characteristics (e.g., physical characteristic machine learning model 122), such as particle packing efficiency and / or rheometry parameters from measured particle characteristics. For example, the machine learning model can include a model that has been trained on experimental data to receive particle characteristics of particles as input, and to generate a predicted output, e.g., an estimate of the chemical composition ratio, particle packing efficiency, an estimate of rheometry7parameters for a concrete mixture, or a combination thereof.

[0147] In some implementations, the machine learning model is a deep learning model that employs multiple layers of models to generate an output for a received input. A deep neural network is a deep machine learning model that includes an output layer and one or more hidden layers that each applies a non-linear transformation to a received input to generate an output. In some cases, the neural network may be a recurrent neural network. A recurrent neural network is a neural network that receives an input sequence and generates an output sequence from the input sequence. In particular, a recurrent neural network uses some or all of the internal state of the network after processing a previous input in the input sequence to generate an output from the current input in the input sequence. In some other implementations, the machine learning model is a convolutional neural network. In some implementations, the machine learning model is an ensemble of models that may' include all or a subset of the architectures described above.

[0148] In some implementations, the machine learning model can be a feedforward autoencoder neural network. For example, the machine learning model can be a three-lay er autoencoder neural network. The machine learning model may include an input layer, ahidden layer, and an output layer. In some implementations, the neural network has no recurrent connections between layers. Each layer of the neural network may be fully connected to the next, there may be no pruning between the layers. The neural network may include an ADAM optimizer, or any other multi-dimensional optimizer, for training the network and computing updated layer weights. In some implementations, the neural network may apply a mathematical transformation, such as a convolutional transformation, to input data prior to feeding the input data to the network.

[0149] In some implementations, the machine learning model can be a supervised model. For example, for each input provided to the model during training, the machine learning model can be instructed as to what the correct output should be. The machine learning model can use batch training, training on a subset of examples before each adjustment, instead of the entire available set of examples. This may improve the efficiency of training the model and may improve the generalizability of the model. The machine learning model may use folded cross-validation. For example, some fraction (the "fold") of the data available for training can be left out of training and used in a later testing phase to confirm how well the model generalizes. In some implementations, the machine learning model may be an unsupervised model. For example, the model may adjust itself based on mathematical distances between examples rather than based on feedback on its performance.

[0150] For the concrete generation system, a machine learning model can be trained to estimate rheometry parameters for concrete mixtures based on measured characteristics of the particles to the mixture. In some examples, the machine learning model can be trained on experimentally determined data relating known characteristics of concrete particles to experimentally determined rheometry parameters. The computing system 402 can then interface with the mixing control 408 or the metering controls system 410 to adjust controls and additives for the carbonation subsystem, the densification subsystem, or both.

[0151] In some examples, the computing system 402 obtains rheometry measurements from the sensors 202 as the concrete mixture is mixed in the carbonation subsystem 124. The system compares the rheometry measurements with estimated rheometry measurements to determine, e.g., whether the concrete mixture will meet desired post-curing mechanical properties or whether additional or additives should be added to the particles.

[0152] For the blast furnace subsystem, the computing system 402 obtains surface area measurements from the sensors, and the computing system 402 can interface with the blast furnace control system 406 or the metering control system 410 to determine adjustments to the blast furnace subsystem 128. The system calculates the SIF for each particle to determine,e.g., whether the particles meet desired properties or whether the computing system should send control signals to adjust the blast furnace subsystem 128 (e.g.. by adjusting an amount of reactions, adjusting a flow rate of the blast air, or adjusting a temperature of the blast air).

[0153] The computing system 402 can include surface area measurement algorithms and one or more lookup tables. For example, in some implementations, the blast furnace control system 406 interfaces with the sensors 202 and receives measurement data 132. The blast furnace control system 406 can process the measurement data 132 to determine particle characteristics of each analyzed particle.

[0154] For example, as discussed in more detail below, the blast furnace control system 406 can execute data analysis algorithms to interpret the measurement data 132 and determine particle characteristics including, but not limited to, surface chemical characteristics and physical characteristics.

[0155] In some implementations, the data analysis algorithms include a mixability model to determine a mixability of the particles based on the particle characteristics. The model can be a theoretical and analytical particle packing model-based Bayesian optimization algorithm-or other machine learning model — to determine a mixability of the particles and estimate quantities of the reactants to add to the blast furnace and control parameters of the blast air.

[0156] In some implementations, the data analysis algorithms include a packing efficiency model to determine a mixability or a packing efficiency of the particles based on the particle characteristics. The model can be a theoretical and analytical particle packing model-based Bayesian optimization algorithm-or other machine learning model — to determine a packing efficiency of the particles and estimate quantities of the reactants to add to the blast furnace and control parameters of the blast air.

[0157] In some implementations, the data analysis algorithms can include a machine learning model to estimate particle mixability or packing efficiency and / or rheometry parameters from measured particle characteristics. For example, the machine learning model can include a model that has been trained on experimental data to receive particle characteristics of particles as input, and to generate a predicted output, e.g.. an estimate of the particle packing efficiency.

[0158] The computing system 402 can use a lookup table 420 of SIFs to correlate measured particle characteristics (e.g., size / shape distributions) to experimentally determined control parameters. For example, the blast furnace control system 406 can compare the measured particle characteristics to entries in the lookup table and estimate the controlparameters based on correlating entries of experimentally determined control parameters in the lookup table. In some examples, the blast furnace control system 406 may interpolate between entries in the lookup table or extrapolate the table data when the measured particle characteristics do not precisely match with a table entry.

[0159] In some implementations, the blast furnace control system 406 can include a machine learning model to estimate an adjustment to a control parameter of the pelletizing drum. In particular, the machine learning model can include a model that has been trained on experimental data to receive particle characteristics, such as surface chemical characteristics or the SIF, as input, and to generate a predicted adjustment to quantities of the reactants to add to the blast furnace, control parameters of the blast air, or both.

[0160] For example, the adjustment to an amount of reactants from measured particle characteristics can be adjusted to an amount of coke to add to the blast furnace. In another example, the predicted adjustment can include, but is not limited to, adjusting the flow rate of the blast air, the temperature of the blast air, or a combination thereof.

[0161] In some implementations, the machine learning model is a deep learning model that employs multiple layers of models to generate an output for a received input. A deep neural network is a deep machine learning model that includes an output layer and one or more hidden layers that each applies anon-linear transformation to a received input to generate an output. In some cases, the neural network may be a recurrent neural network. A recurrent neural network is a neural network that receives an input sequence and generates an output sequence from the input sequence. In particular, a recurrent neural network uses some or all of the internal state of the network after processing a previous input in the input sequence to generate an output from the current input in the input sequence. In some other implementations, the machine learning model is a convolutional neural network. In some implementations, the machine learning model is an ensemble of models that may include all or a subset of the architectures described above.

[0162] A machine learning model can be trained to estimate control parameters for pelletizing particles based on measured characteristics of the particles. In some examples, the machine learning model can be trained on experimentally determined data relating known characteristics of particles to experimentally determined control parameters.

[0163] The computing system 402 can send control signals to the blast furnace based on the determined adjustments to the reactants, the blast air, or both.

[0164] For example, the blast furnace control 406 can interface with the metering control system 410 directly or via the computing system 402 to control the addition of certainquantities of reactants, such as coke or limestone, to the blast furnace. For example, the blast furnace control system 406 can issue commands from the computing system 402 to the metering control system 410 to control an amount of coke to add to the blast furnace based on the chemical composition ratios or the SIFs of the particles, which have a direct impact on the reaction rates of coke and the particles. In particular, the computing system 402 can interface with the metering control system 410 to operate valves from coke and / or limestone supply tanks in order to apply appropriate quantities of coke and / or limestone to the blast furnace.

[0165] In some other examples, the blast furnace control system 406 can interface with the blast furnace subsystem 128 to adjust parameters of the blast air. For example, the blast furnace control system 406 can issue commands from the computing system 402 to the blast furnace to control a temperature of the blast air or a flow rate of the blast air based on the SIFs of the particles, w hich has a direct impact on the w ay that gas moves through the blast furnace. In particular, the computing system 402 can interface with the blast furnace to operate valves from a blast air supply tank.

[0166] FIG. 5 is a flow chart of an example process 500 for outputting relevant characteristics of particles. The process 500 can be performed by the controller 106 discussed above, for example.

[0167] Process 500 can be executed by one or more computing systems including, but not limited to, the controller 106 of material processing system 100 described above, to determine characteristics of particles.

[0168] The system can obtain measurement data (502). For example, the measurement data can include measurement data from multiple sensors positioned along a conveyor belt, directed at different regions of the conveyor belt. The measurement data can include measurement data taken at different times, and / or different batches of particles. The system can obtain measurement data from the sensors.

[0169] In implementations where the sensor rig includes sensors positioned along the convey or belt, such as depicted in FIG. 1 and FIG. 2, each sensor can be directed at different regions of the first section. At a given time, each sensor may be directed at a different batch of particles. Thus, the system can synchronize the measurement data from each sensor with the measurement data of other sensors so the system can match measurement data for a particular batch of particles. At least one of the different types of sensors are positioned to capture measurement data of a top surface of each particle, and the controller is configured to determine a surface area for each of the particles based on the measurement data of the top surface of the particle.

[0170] In some implementations, the sensor rig includes a spray system configured to wet the particles in order to obtain wetted particles.

[0171] In some examples, the system can determine the physical characteristics of the particles based on the measurement data (504). That is, the system can determine the physical characteristics by applying the measurement data to an algorithm, such as a machine learning model trained to determine physical characteristics of particles given the measurement data. The physical characteristics include surface area, volume, composition, shape, water content, specific surface area, or surface roughness. The second machine learning model is trained on training examples comprising measurement data of particles of know n physical characteristics.

[0172] The system can determine, from the measurement data, surface chemical characteristics of the particles based on the measurement data, the physical characteristics, or both (506). In some examples, the system can use an algorithm to determine the surface chemical characteristics based on analyzing and identifying pixels of sensor images of the particles.

[0173] In some examples, the system can apply the measurement data to a machine learning model to determine surface chemical characteristics of particles given measurement data. The system can apply the measurement data to a machine learning model configured to determine surface chemical characteristics of particles given measurement data as described above with reference to FIG. 2. The measurement data can include images of the particles, such as thermal images, infrared images, or both.

[0174] In some implementations, the measurement data that is provided to the machine learning model is synchronized measurement data for the same batch of particles, For example, the system can provide the output of the model that synchronizes measurement data as the input to the machine learning model configured to determine surface chemical characteristics of particles given measurement data. The machine learning model can be trained on training examples comprising measurement data of particles of known surface chemical characteristics.

[0175] In particular, the machine learning model can determine the surface chemical characteristics by identifying regions of pixels of the images corresponding to the particles. The surface chemical characteristics of the particles include an oxidation rate, a reactivity rate, a chemical composition ratio, or a moisture evaporation rate. For example, the system can use machine learning to determine the surface chemical characteristics by identifying one or more chemicals on the surface of the particles based on the identified pixels, and themachine learning model can generate a chemical composition ratio for the particles. The chemical composition ratio is a percentage content of each of the one or more chemicals.

[0176] In another example, the machine learning model can determine a moisture evaporation rate of wetted particles based on measurement data including measurements of the surface of the wetted particles.

[0177] In some implementations, the system can apply the measurement data to a second machine learning model trained to determine physical characteristics of particles given the measurement data. The physical characteristics include surface area, volume, composition, shape, water content, specific surface area, or surface roughness. The second machine learning model is trained on training examples comprising measurement data of particles of known physical characteristics.

[0178] The system can output the surface chemical characteristics of the particles from the machine learning model to one or more components along the conveyor belt (506). The system can output relevant characteristics of particles from the machine learning model to one or more components along the conveyor belt. As described above with reference to FIG. 2, different characteristics may be relevant for different components along the conveyor belt. For example, the surface area may be relevant for a component that is a measuring device, while the particle shape may be relevant for a component that is a mixing vessel. In order to output relevant characteristics to each component, the system can determine which of the characteristics outputted from the machine learning model is relevant to each of the components.

[0179] In some implementations, the system can output the physical characteristics of the particles from the second machine learning model to one or more components along the conveyor belt.

[0180] In some implementations, the system can adjust control settings of a chemical processing subsystem based on the surface chemical characteristics, the physical characteristics, or both. The chemical processing subsystem can be a blast furnace subsystem, a carbonation subsystem, or a densification subsystem. In particular, the system can determine the adjustments to control settings of a particle processing system based at least in part on the surface chemical characteristics, the physical characteristics, or both. The adjustments, in terms of a carbonation and densification process, can include adjusting an amount of additives, an adjustment to a spin speed, an adjustment to a feed rate, or a combination thereof. In some examples, in terms of a blast furnace smelting process, theadjustments include adjustment to a flow rate of the example particle processing system, a temperature of the example particle processing system, or a combination thereof.

[0181] In some examples, the system can determine the adjustments to the control settings by calculating a reactivity of the particles based on the surface chemical characteristics, the physical characteristics, or both, and the system can determine the adjustments based on the calculated reactivity.

[0182] The system can thus provide relevant and up-to-date information to downstream processing systems to make decisions or adjustments. For example, as described above with reference to FIG. 2, downstream processing systems can use the information to make decisions such as concrete mix recipes, adjustments to mining direction, or tool settings.

[0183] FIG. 6 depicts a schematic diagram of a computer system that may be applied to any of the computer-implemented methods and other techniques described herein. The system 600 can be used to carry out the operations described in association with any of the computer-implemented methods described previously, according to some implementations. In some implementations, computing systems and devices and the functional operations described in this specification can be implemented in digital electronic circuitry, in tangibly- embodied computer software or firmware, in computer hardware, including the structures disclosed in this specification (e.g., system 600) and their structural equivalents, or in combinations of one or more of them. The system 600 is intended to include various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers, including vehicles installed on base units or pod units of modular vehicles. The system 600 can also include mobile devices, such as personal digital assistants, cellular telephones, smartphones, and other similar computing devices. Additionally, the system can include portable storage media, such as Universal Serial Bus (USB) flash drives. For example, the USB flash drives may store operating systems and other applications. The USB flash drives can include input / output components, such as a wireless transducer or USB connector that may be inserted into a USB port of another computing device.

[0184] The system 600 includes a processor 610, a memory 620, a storage device 630. and an input / output device 640. Each of the components 610, 620, 630, and 640 are interconnected using a system bus 650. The processor 610 is capable of processing instructions for execution within the system 600. The processor may be designed using any of a number of architectures. For example, the processor 610 may be a CISC (ComplexInstruction Set Computers) processor, a RISC (Reduced Instruction Set Computer) processor, or a MISC (Minimal Instruction Set Computer) processor.

[0185] In one implementation, the processor 610 is a single-threaded processor. In another implementation, the processor 610 is a multi-threaded processor. The processor 610 is capable of processing instructions stored in the memory 620 or on the storage device 630 to display graphical information for a user interface on the input / output device 640.

[0186] The memory 620 stores information within the system 600. In one implementation, the memory 620 is a computer-readable medium. In one implementation, the memory 620 is a volatile memoiy unit. In another implementation, the memory 620 is a nonvolatile memory unit.

[0187] The storage device 630 is capable of providing mass storage for the system 600. In one implementation, the storage device 630 is a computer-readable medium. In various different implementations, the storage device 630 may be a floppy disk device, a hard disk device, an optical disk device, or a tape device.

[0188] The input / output device 640 provides input / output operations for the system 600. In one implementation, the input / output device 640 includes a keyboard and / or pointing device. In another implementation, the input / output device 640 includes a display unit for displaying graphical user interfaces.

[0189] Implementations of the subject matter and the functional operations described in this specification can be implemented in digital electronic circuitry, in tangibly -implemented computer software or firmware, in computer hardware, including the structures disclosed in this specification and their structural equivalents, or in combinations of one or more of them. Implementations of the subject matter described in this specification can be implemented as one or more computer programs, i.e., one or more modules of computer program instructions encoded on a tangible non- transitory program earner for execution by, or to control the operation of, data processing apparatus. The computer storage medium can be a machine- readable storage device, a machine-readable storage substrate, a random or serial access memory device, or a combination of one or more of them.

[0190] The term "data processing apparatus7’ refers to data processing hardware and encompasses all kinds of apparatus, devices, and machines for processing data, including, by way of example, a programmable processor, a computer, or multiple processors or computers. The apparatus can also be or further include special purpose logic circuitry, e.g., a central processing unit (CPU), a FPGA (field programmable gate array), or an ASIC (applicationspecific integrated circuit). In some implementations, the data processing apparatus and / orspecial purpose logic circuitry may be hardware-based and / or software-based. The apparatus can optionally include code that creates an execution environment for computer programs, e.g., code that constitutes processor firmware, a protocol stack, a database management system, an operating system, or a combination of one or more of them. The present disclosure contemplates the use of data processing apparatuses with or without conventional operating systems, for example Linux, UNIX, Windows, Mac OS, Android, iOS or any other suitable conventional operating system.

[0191] A computer program, which may also be referred to or described as a program, software, a software application, a module, a software module, a script, or code, can be written in any form of programming language, including compiled or interpreted languages, or declarative or procedural languages, and it can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment. A computer program may, but need not, correspond to a file in a file system. A program can be stored in a portion of a file that holds other programs or data, e.g., one or more scripts stored in a markup language document, in a single file dedicated to the program in question, or in multiple coordinated files, e.g.. files that store one or more modules, sub-programs, or portions of code. A computer program can be deployed to be executed on one computer or on multiple computers that are located at one site or distributed across multiple sites and interconnected by a communication network. While portions of the programs illustrated in the various figures are shown as individual modules that implement the various features and functionality through various objects, methods, or other processes, the programs may instead include a number of sub-modules, third party services, components, libraries, and such, as appropriate. Conversely, the features and functionality of various components can be combined into single components as appropriate.

[0192] The processes and logic flows described in this specification can be performed by one or more programmable computers executing one or more computer programs to perform functions by operating on input data and generating output. The processes and logic flows can also be performed by, and apparatus can also be implemented as, special purpose logic circuitry, e.g., a central processing unit (CPU), a FPGA (field programmable gate array), or an ASIC (application-specific integrated circuit).

[0193] Computers suitable for the execution of a computer program include, by way of example, can be based on general or special purpose microprocessors or both, or any other kind of central processing unit. Generally, a central processing unit will receive instructions and data from a read-only memory or a random access memory or both. The essentialelements of a computer are a central processing unit for performing or executing instructions and one or more memory’ devices for storing instructions and data. Generally, a computer will also include, or be operatively coupled to receive data from or transfer data to, or both, one or more mass storage devices for storing data, e.g., magnetic, magneto-optical disks, or optical disks. However, a computer need not have such devices. Moreover, a computer can be embedded in another device, e.g., a mobile telephone, a personal digital assistant (PDA), a mobile audio or video player, a game console, a Global Positioning System (GPS) receiver, or a portable storage device, e.g., a universal serial bus (USB) flash drive, to name just a few.

[0194] Computer-readable media (transitory’ or non-transitory, as appropriate) suitable for storing computer program instructions and data include all forms of non-volatile memory', media and memory devices, including by way of example semiconductor memory devices, e.g., EPROM, EEPROM, and flash memory devices; magnetic disks, e.g., internal hard disks or removable disks; magneto-optical disks; and CD-ROM and DVD-ROM disks. The memory’ may store various objects or data, including caches, classes, frameworks, applications, backup data, jobs, web pages, web page templates, database tables, repositories storing business and / or dynamic information, and any other appropriate information including any parameters, variables, algorithms, instructions, rules, constraints, or references thereto. Additionally, the memory’ may include any other appropriate data, such as logs, policies, security or access data, reporting files, as well as others. The processor and the memory can be supplemented by, or incorporated in, special purpose logic circuitry.

[0195] To provide for interaction with a user, implementations of the subject matter described in this specification can be implemented on a computer having a display device, e.g., a CRT (cathode ray tube), LCD (liquid crystal display), or plasma monitor, for displaying information to the user and a keyboard and a pointing device, e.g., a mouse or a trackball, by which the user can provide input to the computer. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory' feedback, e.g., visual feedback, auditory' feedback, or tactile feedback; and input from the user can be received in any form, including acoustic, speech, or tactile input. In addition, a computer can interact with a user by sending documents to and receiving documents from a device that is used by the user; for example, by sending web pages to a web browser on a user’s client device in response to requests received from the web browser.

[0196] The term “graphical user interface.” or GUI. may be used in the singular or the plural to describe one or more graphical user interfaces and each of the displays of aparticular graphical user interface. Therefore, a GUI may represent any graphical user interface, including but not limited to, a web browser, a touch screen, or a command line interface (CLI) that processes information and efficiently presents the information results to the user. In general, a GUI may include a plurality of user interface (UI) elements, some or all associated with a web browser, such as interactive fields, pull-down lists, and buttons operable by the business suite user. These and other UI elements may be related to or represent the functions of the web browser.

[0197] Implementations of the subject matter described in this specification can be implemented in a computing system that includes a back-end component, e.g., as a data server, or that includes a middleware component, e.g., an application server, or that includes a front-end component, e.g.. a client computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the subject matter described in this specification, or any combination of one or more such back-end, middleware, or front-end components. The components of the system can be interconnected by any form or medium of digital data communication, e.g., a communication network. Examples of communication networks include a local area network (LAN), a wide area network (WAN), e.g., the Internet, and a wireless local area network (WLAN).

[0198] The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.

[0199] While this specification contains many specific implementation details, these should not be construed as limitations on the scope of any invention or on the scope of what may be claimed, but rather as descriptions of features that may be specific to particular implementations of particular inventions. Certain features that are described in this specification in the context of separate implementations can also be implemented in combination in a single implementation. Conversely, various features that are described in the context of a single implementation can also be implemented in multiple implementations separately or in any suitable sub-combination. Moreover, although features may be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claimed combination may be directed to a sub-combination or variation of subcombinations.

[0200] Similarly, while operations are depicted in the drawings in a particular order, this should not be understood as requiring that such operations be performed in the particular order shown or in sequential order, or that all illustrated operations be performed, to achieve desirable results. In certain circumstances, multitasking and parallel processing may be helpful. Moreover, the separation of various system modules and components in the implementations described above should not be understood as requiring such separation in all implementations, and it should be understood that the described program components and systems can generally be integrated together in a single software product or packaged into multiple software products.

[0201] Particular implementations of the subject matter have been described. Other implementations, alterations, and permutations of the described implementations are within the scope of the following claims as will be apparent to those skilled in the art. For example, the actions recited in the claims can be performed in a different order and still achieve desirable results.

[0202] Accordingly, the above description of example implementations does not define or constrain this disclosure. Other changes, substitutions, and alterations are also possible without departing from the spirit and scope of this disclosure.

[0203] What is claimed is:

Claims

CLAIMS1. A particle characterization system comprising: a conveyor belt comprising a first section having one or more mechanisms arranged to create environmental conditions favorable to sensor readings; a sensor rig located proximate to the first section of the conveyor belt comprising a plurality of different types of sensors arranged to obtain measurements of particles conveyed by the first section of the conveyor belt; and a controller configured to: obtain measurement data from the plurality of different types of sensors; determining, from the measurement data, surface chemical characteristics of the particles based on the measurement data; and output the surface chemical characteristics of the particles from the machine learning model to one or more components along the conveyor belt.

2. The system of claim 1. wherein the surface chemical characteristics of particles comprise any one of: an oxidation rate, a reactivity rate, a chemical composition ratio, or a moisture evaporation rate.

3. The system of claim 1. wherein determining, from the measurement data, surface chemical characteristics of the particles based on the measurement data comprises: applying the measurement data to a machine learning model trained to determine the surface chemical characteristics.

4. The system of claim 3, wherein the machine learning model is trained using training data comprising measurement data of particles of known surface chemical characteristics.

5. The system of claim 2, wherein the measurement data comprises at least one of thermal images of the particles, infrared images of the particles, or hyper-spectral images of the particles.

6. The system of claim 4, wherein determining, from the measurement data, surface chemical characteristics of the particles based on the measurement data comprises:identifying regions of pixels of the images corresponding to the particles.

7. The system of claim 6, wherein determining the surface chemical characteristics of the particles comprises: identifying one or more chemicals present on the surface of the particles based on applying the regions of pixels as input to a machine learning model trained to identify surface chemical opposition from image data; and generating, based on output from the machine learning model, a chemical composition ratio for the particles, the chemical composition representing a percentage content of each of the one or more chemicals.

8. The system of claim 1, wherein the plurality of different types of sensors are positioned along the conveyor belt to point at different regions of the conveyor belt, and the controller is further configured to synchronize measurement data from the sensors.

9. The system of claim 8, wherein the plurality of sensors comprises optical sensors, infrared sensors, or a combination thereof.

10. The system of claim 9. wherein at least one of the plurality of different types of sensors is positioned to capture measurement data of a top surface of the particle, and wherein the controller is configured to determine a surface area for each of the particles based on the measurement data of the top surface of the particle.

11. The system of claim 9, wherein the sensor rig comprises a spray system configured to wet the particles to obtain wetted particles prior to obtaining the measurement data of the wetted particles from an infrared sensor of the plurality of sensors.

12. The system of claim 11, wherein the system determines a moisture evaporation rate at the surface of the wetted particles based on obtaining the measurement data of the wetted particles.

13. The system of claim 3. wherein the controller is further configured to:apply the measurement data to a second machine learning model trained to determine physical characteristics of particles given the measurement data.

14. The system of claim 13, wherein the physical characteristics of particles comprise any one of: surface area, volume, composition, shape, water content, specific surface area, or surface roughness.

15. The system of claim 13, wherein the second machine learning model is trained on training examples comprising measurement data of particles of known physical characteristics.

16. The system of claim 13, wherein the physical characteristics are outputted to different components along the conveyor belt.

17. The system of claim 13, further comprising: adjusting control settings of a chemical processing subsystem based at least in part on the surface chemical characteristics, the physical characteristics, or both.

18. The system of claim 13, further comprising: determining adjustments to control settings of a particle processing system based at least in part on the surface chemical characteristics, the physical characteristics, or both.

19. The system of claim 17, wherein determining adjustments to control settings of the chemical processing subsystem further comprises: adjusting an amount of additives, an adjustment to a spin speed, an adjustment to a feed rate, or a combination thereof.

20. The system of claim 18, wherein determining adjustments to control settings of the particle processing system further comprises: calculating a reactivity of the particles based at least in part on the surface chemical characteristics, the physical characteristics, or both; and determining adjustments to the control settings based on the reactivity.

21. The system of claim 20, wherein the adjustments to the control settings comprise: an adjustment to a flow rate of the particle processing system, a temperature of the particle processing system, or a combination thereof.

22. A method for determining characteristics of particles comprising: obtaining, by one or more processors, measurement data from a plurality of different types of sensors arranged relative to a particular section of a conveyor belt to obtain measurements of particles conveyed along the particular section of the conveyor belt, and wherein the particular section of the conveyor belt comprises one or more mechanisms arranged to create environmental conditions favorable to sensor reading; determining, from the measurement data, surface chemical characteristics of the particles based on the measurement data; and outputting, by the one or more processors, the surface chemical characteristics of particles from the machine learning model to one or more components along the conveyor belt.

23. A non-transitory computer storage medium encoded with instructions that, when executed by one or more computers, cause the one or more computers to perform operations comprising: obtaining, by one or more processors, measurement data from a plurality of different types of sensors arranged relative to a particular section of a conveyor belt to obtain measurements of particles conveyed along the particular section of the conveyor belt, and wherein the particular section of the conveyor belt comprises one or more mechanisms arranged to create environmental conditions favorable to sensor reading; determining, from the measurement data, surface chemical characteristics of the particles based on the measurement data; and outputting, by the one or more processors, the surface chemical characteristics of particles from the machine learning model to one or more components along the conveyor belt.

Images