Ore sample online sorting system based on mineralogical analysis

Abstract

An online ore sample sorting system based on mineralogical analysis includes a conveyor moving a stream of ore samples along a transport path, a LIBS module, a height measurement device, a focusing controller, and an airflow system. The LIBS module projects a LIBS laser beam along an optical path, the LIBS laser beam is focused on an analysis spot, the analysis spot has a size on the order of magnitude of a size of one or several mineral components of the ore sample, and the LIBS module collects a returned LIBS optical signal. The height measurement device and the focusing controller cooperate to adjust a focus of the LIBS laser beam in real time to move the analysis spot vertically with respect to the conveyor depending on a height of the ore sample passing through the optical path. The airflow system moves particles away from the optical path. A processing unit performs a mineralogical analysis of the ore sample based on the LIBS optical signal.

Description

Technical Field

[0001] This technical field generally relates to the field assessment of ore samples, and more specifically to systems that use LIBS to obtain mineralogical information of ore samples in real time. Background Technology

[0002] To meet the growing global demand for critical metals, the mining industry faces the challenge of adopting more intelligent approaches to adapt to declining ore grades and increasing environmental pressures. Successful, environmentally conscious exploration and mineral extraction require effective analysis of the elemental composition and mineralogical properties of ore-bearing rocks.

[0003] Mineral characterization is a complex task that may involve measuring elemental composition, mineralogical properties, lithology, hardness, and various other chemical or physical parameters. In the context of mining operations, mining companies typically need to understand ore characteristics for mining planning and operations. Among other parameters, ore mineralogical properties significantly impact the cost and efficiency of metal extraction. The metallurgical properties of the ore also influence its characteristics during processing steps such as crushing, concentration, and extraction.

[0004] Raw materials have a significant impact on the quality of the final refractory product because the ore used for processing often varies considerably in chemical and mineral composition. Therefore, mining operators seek to ensure consistent raw material quality using pre-defined quality factors. This can be achieved by separating raw materials by grade (e.g., removing portions unsuitable for specific applications) and by rationally controlling processing parameters based on real-time information about the chemical composition and mineralogical properties of the raw materials. To achieve this separation, rapid online technology is required, preferably providing synchronous, real-time data on the elemental and mineralogical composition of the raw material or ore-bearing rock.

[0005] Several techniques for ore analysis are known in the art, using data from multiple sensors, each dedicated to a specific result. However, each of these techniques is best suited for somewhat different lists of mineral species.

[0006] Near-infrared (NIR) analysis is generally highly useful for measuring altered minerals. These minerals are the result of alteration of the host rock and are often precursors to mineralization (the presence of valuable metals). While NIR is sensitive to most altered minerals, it is less sensitive to many rock-forming and sulfide minerals.

[0007] FT-IR (Fourier Transform Infrared) analysis is generally useful for measuring rock-forming minerals, but less so for alteration minerals. However, FT-IR is not suitable for metamorphic minerals.

[0008] Element determination.

[0009] Raman analysis is useful for sulfides, crystalline materials, and some rock-forming minerals, but it is insufficient for elemental determination. Raman analysis systems use laser beams of different wavelengths to excite the atoms in a sample, causing them to enter different vibrational states. Some of these vibrational states result in energy changes in certain portions of the incident illumination. Raman measures the transfer of energy states, and materials typically possess a unique fingerprint for Raman transfers.

[0010] The combination of NIR, FT-IR, and Raman sensors helps produce better qualitative and quantitative results. However, all of these techniques require an additional elemental analysis method. In fact, knowledge of elemental composition can aid in the identification of these minerals, as NIR, FT-IR, and Raman are purely molecular techniques. Therefore, additional techniques are often needed to measure the elemental composition of minerals, such as X-ray fluorescence spectroscopy (XRF) or various chemical analyses. Furthermore, the NIR region is not ideal for measuring many rock-forming minerals, so Raman or Fourier transform infrared (FT-IR) spectroscopy is often used alone. Additional chemical or physical tests can also be used to provide measurements of the metallurgical processing parameters of the material.

[0011] Quantitative mineral analysis (QMA) using energy-dispersive X-ray spectroscopy and scanning electron microscopy (EDS-SEM) provides reliable information on the mineral abundance and structure of prepared rocks. However, electron microscopy-based instruments are designed for laboratory use and require time-consuming sample preparation (polishing and sputtering of carbon), making them inconvenient offline techniques for rapid on-site measurements.

[0012] Laser-induced breakdown spectroscopy (LIBS) has been used for elemental analysis in many environments and has recently been demonstrated for mineral quantification and identification. El Haddad et al. (Multiphase mineral identification and quantification by LIBS, Minerals Engineering, Vol. 134, pp. 281-290) presented a novel laboratory method for mineral identification and quantification using LIBS that is scalable to perform automated mineralogical measurements in coarse-grained rocks in a faster manner than conventional QMA-based methods.

[0013] While any given analytical technique may be able to provide a subset of the required information, in many cases, the accuracy and precision of that single analytical technique may not be optimal for online analysis required for ore mineralogical purposes or for elemental analysis needed to improve mining efficiency through ore sorting or feed monitoring. Using sensor-based ore sorting can significantly increase the efficiency of the process and substantially improve yield.

[0014] There is still a need for online mineralogical analysis of ore samples, rather than using the traditional method of extracting samples for offline analysis in the laboratory, which is expensive and time-consuming and not suitable for rapid analysis that can sort samples. Summary of the Invention

[0015] According to one aspect, a system for online sorting of ore samples based on mineralogical analysis is provided, comprising:

[0016] - A conveyor for moving a stream of ore samples along a conveying path, the ore samples having a variable sample height on the conveyor;

[0017] - LIBS module, which projects a pulsed LIBS laser beam along the optical path, and focuses the pulsed LIBS laser beam onto the analysis spot on the transport path. The size of the analysis spot is on the order of magnitude of the size of one or more mineral components in the ore sample. The LIBS module collects the LIBS light signal returning along the optical path.

[0018] - A height measuring device configured to measure in real time the sample height of the top surface of the ore sample stream along a conveyor travel axis intersecting the LIBS laser beam at a point upstream of the LIBS module.

[0019] - A focusing controller, which communicates with the height measuring device and is configured to focus the LIBS laser beam at the sample height in sync with the movement of the ore sample and the optical pulse;

[0020] - An airflow system configured to generate at least one airflow that disperses particles away from the optical path; and

[0021] - A processing unit, which performs mineralogical analysis on the ore sample based on the LIBS optical signal.

[0022] In some embodiments, the LIBS laser beam has a diameter of about 70 μm to about 140 μm at the analysis spot, and preferably has a diameter of about 100 μm.

[0023] In some embodiments, the LIBS module includes a focusing lens and a translational lens holder, the focusing lens focusing the LIBS laser beam at the analysis spot, and the translational lens holder being configured to move the focusing lens vertically under the control of the focusing controller.

[0024] In some embodiments, the system further includes a conveyor speed measuring mechanism configured to provide real-time measurement of the travel speed of the ore sample on the conveyor and operatively connected to the focusing controller.

[0025] In some embodiments, the conveyor speed measuring mechanism may include a rotary encoder in contact with a portion of the conveyor, the rotary encoder rotating at a rotational speed matching the travel speed of the ore sample on the conveyor. In other embodiments, the conveyor speed measuring mechanism includes a first distance sensor and a second distance sensor, the first distance sensor and the second distance sensor being located above the conveyor, on the same plane parallel to the surface of the conveyor, the first distance sensor and the second distance sensor being spaced apart along the conveyor's travel axis by a predetermined spacing. The second distance sensor is collinearly aligned with the LIBS laser beam in a vertical direction. The focusing controller includes an FPGA configured to sample a variable sample height signal from the height measuring device at a sampling frequency higher than the laser repetition rate of the LIBS laser beam. The FPGA includes an analog-to-digital converter.

[0026] In some embodiments, the airflow system includes a main nozzle mounted between the LIBS module and the delivery path. The main nozzle may have an upper end and a lower end that allow light propagation of the LIBS laser beam and plasma light, and the main nozzle is shaped as a frustoconical shape that tapers gradually from the upper end to the lower end. The airflow system may include a main blower unit that generates a escort airflow and is connected to the main nozzle to inject the escort airflow into the main nozzle near the upper end. The upper end of the main nozzle is closed by a top wall to prevent airflow but allow light to pass through.

[0027] In some embodiments, the airflow system includes an auxiliary nozzle adjacent to the main nozzle and oriented at a small angle to the optical path, the auxiliary nozzle generating a clean airflow toward the area where the optical path intersects the delivery path.

[0028] The airflow system may include an auxiliary blower unit that generates the clean airflow and is connected to the auxiliary nozzle.

[0029] In some embodiments, the airflow system includes a scraper nozzle disposed upstream of the LIBS module and above the transport path, and the scraper nozzle is configured and shaped to generate a sufficiently strong scraper airflow to remove unwanted material from the surface of the ore sample.

[0030] In some embodiments, the system includes one or more protective mechanisms to prevent ore samples on the transport path from damaging components of the airflow system.

[0031] In some embodiments, the mineralogical analysis performed by the processor includes identifying and quantifying individual mineral features of the composition of the ore sample using chemometric data processing methods for mixed-spectral deconvolution.

[0032] Other features and advantages will be better understood when reading the preferred embodiments with reference to the accompanying drawings. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of a system according to one implementation method.

[0034] Figure 2 This is a schematic diagram of a LIBS module according to one implementation method.

[0035] Figure 3A and 3B This is a schematic diagram of variations of autofocus based on different implementation methods.

[0036] Figure 4 This is a schematic diagram of an airflow system according to one embodiment.

[0037] Figure 5A and 5B The illustration shows a method according to one embodiment. Figure 1 The system shown illustrates the use of protection mechanisms.

[0038] Figure 6A and 6B These are images comparing QMA and LIBS mappings.

[0039] Figure 7A and 7B The results of mineral phase abundance predictions and the associated absolute errors for various minerals are shown. Detailed Implementation

[0040] This description relates to a system for online sorting of ore samples based on mineralogical analysis of these ore samples.

[0041] In the following description, similar features in the accompanying drawings are given similar reference numerals. To avoid unduly obscuring the drawings, elements that have already been mentioned in the preceding drawings may not be labeled in some of the drawings. It should also be understood that the elements in the drawings are not necessarily drawn to scale, but rather the emphasis is on clearly illustrating the elements and structures of this embodiment.

[0042] The terms “a,” “an,” and “one” are defined herein as meaning “at least one,” meaning that, unless otherwise stated, these terms do not exclude multiple items. Terms modifying the values, conditions, or characteristics of features of exemplary embodiments, such as “generally,” “approximately,” and “about,” should be understood to mean that the value, condition, or characteristic is defined within acceptable tolerances for proper operation of the exemplary embodiment for its intended application. For example, the term “about” may mean within an acceptable range of error for a particular value as determined by one of ordinary skill in the art, which will depend in part on how the value is measured or determined, i.e., the limitations of the measurement system. Measurements with 10%–20% accuracy are generally considered acceptable and include the term “about.”

[0043] In this description, when a wide range of numerical values ​​is provided, any possible narrower ranges within the boundaries of the wider range are also considered. For example, if a wide range of values ​​from 0 to 1000 is provided, any narrower range between 0 and 1000 is also considered. If a wide range of values ​​from 0 to 1 is mentioned, any narrower range between 0 and 1, i.e., decimal values, is also considered.

[0044] Unless otherwise stated, the terms “connected” and “coupled” and their derivatives and variations herein refer to any structural or functional connection or coupling, directly or indirectly, between two or more elements. For example, a connection or coupling between elements can be mechanical, optical, electrical, logical, or any combination thereof.

[0045] In this description, the terms "light" and "optics," and their variations and derivatives, are used to refer to radiation in any suitable region of the electromagnetic spectrum. Therefore, the terms "light" and "optics" are not limited to visible light, but may also include, but are not limited to, the infrared or ultraviolet regions of the electromagnetic spectrum. Furthermore, those skilled in the art will understand that the definitions of the ultraviolet, visible, and infrared ranges in terms of spectral range, and the boundaries between them, may vary depending on the technical field or definition considered, and are not intended to limit the scope of application of this technology.

[0046] In various implementations, ore samples can be from mining sources in a general sense, containing natural materials derived from the surface. Mining samples and surface-extracted samples typically include at least one valuable ore mineral species mixed with gangue composed of unwanted or worthless rocks and minerals, as well as non-mineral species such as organic materials, bitumen, etc. The systems and methods described herein can be used to sort ore samples containing a mixture of valuable and worthless mineral species (also referred to herein as mineral contaminants). Ore feedstocks with high mineral contaminant loads are often processed, but these processes are expensive and / or have long-term environmental costs. Classifying ore samples according to their mineral content allows for the separation of valuable from worthless samples. Sorting thresholds can be defined by the operator.

[0047] Ore sorting has the potential to improve the range of ores that processing plants can efficiently handle, thereby significantly reducing downstream operating costs and limiting environmental hazards by reducing mine waste, improving ore quality, and increasing mineral recovery rates. Ore sorting technologies can remove gangue and low-grade ores before the main processing steps. Overall energy, material, and labor costs may be reduced, while a significant amount of more valuable minerals are separated from the waste.

[0048] system

[0049] Figure 1 A system 20 for online sorting of ore samples based on mineralogical analysis, according to one embodiment, is schematically illustrated.

[0050] System 20 includes a conveyor 22 for moving a stream of ore samples 24 along a conveying path 26. For example, conveyor 22 may be implemented by a conveyor typically used to transport ore samples at a mining site.

[0051] LIBS module

[0052] System 20 also includes a LIBS module 30 for projecting a pulsed LIBS laser beam 32 along the optical path 35 and collecting the LIBS optical signal 42 returning along the optical path 35. The pulsed LIBS laser beam 32 is focused on an analysis spot 36 and scans over the variable-height contents of the transport path 26.

[0053] The abbreviation LIBS is well-known in the field and stands for Laser-Induced Breakdown Spectroscopy (or spectrometry). LIBS is a well-known technique for extracting elemental information from a given sample. A typical LIBS measurement is performed as follows: a short laser pulse is sent and focused onto the sample surface; the surface is rapidly heated by the laser pulse, causing partial material evaporation and the conversion of gas into plasma, the composition of which represents the elemental content of the sample; as the plasma cools, the excited electrons in the plasma eventually return to the ground state of their associated atoms, and the radiating electrons recombine and emit photons with discrete energies allowed by their associated atomic energy levels; the emitted photons are collected and sent to a spectrometer to produce a spectrum. The spectral distribution (intensity and frequency) of the collected plasma light is correlated with the elemental composition of the plasma, thus determining the elemental composition of the sample. LIBS provides a rapid, localized, non-contact, and sensitive measurement of the elemental composition of materials. LIBS is associated with identifying major and trace elements, the latter typically measured with a sensitivity of parts per million (ppm). Some advantages of LIBS include its ability to be implemented under ambient conditions and its capacity to acquire spectra from targets at considerable distances without any separate sample preparation (e.g., no chemical or solvent-based sample preparation, or a laser ablation process using only the same or a different laser as the LIBS laser). Other advantages of LIBS include its ability to detect low-level (ppm) components, including lighter elements such as Be, B, Li, Na, C, and F. From a practical standpoint, implementing LIBS can significantly reduce latency in sample preparation and data acquisition, enabling real-time (or near-real-time) decision-making.

[0054] Figure 2The illustration schematically depicts a configuration of an example LIBS module 30. In the illustrated embodiment, the LIBS module 30 includes a pulsed laser source 34. The pulsed laser source 34 is configured to emit a pulsed beam containing a LIBS laser beam 32, which travels along the optical path 35 leading to the ore sample 24 as the ore sample 24 travels along its own journey along the transport path 26. The LIBS laser beam 32 has a flux suitable for evaporating a volume of sample 24 at the analytical spot 36 to generate a plasma 38 of material. In some embodiments, the pulsed laser source 34 may operate in different settings (e.g., flux, duty cycle, pulse duration, repetition rate) depending on its intended use, such as in a first set of settings for performing LIBS measurements and in a second set of settings for performing laser cleaning. For example, the pulsed laser source 34 may be implemented by a pulsed Nd:YAG laser source that generates laser pulses with a wavelength of 1064 nm. Pulse energy and duration can be specifically varied depending on the task at hand: cleaning material removal; or the LIBS measurement itself. Pulse energy can range from a few microjoules to hundreds of millijoules. Pulse duration can range from a few femtoseconds to hundreds of nanoseconds. Beam intensity typically reaches GW / cm at 36 of the analytical spot. 2 Status. The laser repetition rate depends on the laser source parameters, typically ranging from a few hertz to hundreds of kilohertz.

[0055] The LIBS module 30 may also include a spectrally resolved photodetector 40 configured to detect light from the plasma, referred to herein as “plasma light” or “LIBS light signal.” The spectrally resolved photodetector 40 may include, for example, optics, mirrors, and one or more spectrometers. The spectrometer is selected based on the required measurement needs. Key parameters of the spectrometer are, but are not limited to, its luminous flux, its sensitivity, its spectral range, its spectral resolution, and its ability to gate measurements in a timely manner.

[0056] LIBS module 30 may also include LIBS controller 31. The LIBS controller may be implemented by one or more control devices that provide the required operating functions of LIBS module 30, such as a laser control processor, circuit, or processor (software-defined) module for providing drive signals to pulsed laser source 34, a spectrometer processor, circuit, or module for operating the spectrometer (including controlling any moving parts or imaging elements), and an output control processor, circuit, or processor module for receiving data from the spectrometer device, digitizing the data, extracting feedback from it, and outputting calculated or measured values.

[0057] The system 20 described herein may also include any number of beam-directing optics that collectively direct the LIBS laser beam 32 from the pulsed laser source 34 to the analysis spot 36, and the plasma light 42 from the plasma 38 to the spectral-resolved detector 40. In the illustrated embodiment, by way of example only, the LIBS module 30 also includes a first lens 44 and a dichroic plate 46 sequentially along the path of the LIBS laser beam 32. The LIBS laser beam 32 is guided and focused onto the analysis spot 36. The laser pulse evaporates and ionizes a portion of the sample 24 at the analysis spot 36 to form plasma 38. The plasma light 42 is reflected from the dichroic plate 46 and focused onto the spectral-resolved detector 40 by the second lens 48.

[0058] The LIBS laser beam 32 is focused at a location referred to herein as analytical spot 36. As explained further below, the focus of the LIBS laser beam 32 is adjustable relative to the transport path 26 such that analytical spot 36 coincides with the top surface of the ore sample 24 directly below the LIBS module 30. In some embodiments, a focusing lens 25 is provided as an output of the LIBS module 30 and may or may not be integrated with the LIBS module 30. The focusing lens 25 preferably has a vertically adjustable position to provide focus adjustment and thus move analytical spot 36. The interaction between the LIBS laser beam 32 and the material of the ore sample 24 at analytical spot 36 results in the generation of plasma 38, which is generated by ionizing the elemental composition of the material of the ore sample 24 at analytical spot 36. The LIBS module is configured such that the size of analytical spot 36 is on the order of magnitude of the size of one or more mineral compositions of the ore sample 24. For example, in some embodiments, the LIBS laser beam 32 has a beam diameter on the order of 100 μm at its focal point, and therefore at the analytical spot. In some embodiments, the beam diameter of the LIBS laser beam 32 at its focal point is between about 70 μm and about 140 μm. In this way, the ionized material in the plasma in each measurement has a relative concentration reflecting the composition of a particular mineral surface, and is a more unique identifier in materials mineralogy.

[0059] Plasma 38 emits plasma light 42, which has a spectral content corresponding to the common emission spectrum of different ionized substances in plasma 38, weighted universally and subject to absorption and loss. The LIBS module collects the plasma light 42 as it propagates along the reverse optical path 35 of the LIBS laser beam 32 toward the LIBS module 30. The LIBS light signal is ultimately detected by a spectrally resolved photodetector 40, providing a spectrally resolved detector signal, which is an electrical signal representing the intensity of the collected light according to wavelength. The positions and relative intensities of the different emission lines of the spectrum associated with each element of the plasma are known from literature and / or calibration data. A processor 50 is provided to communicate with the spectrally resolved photodetector 40 for performing mineralogical analysis of ore samples based on the LIBS light signal, as explained in further detail below.

[0060] To ensure accurate assessment of the mineralogical content of ore samples, the LIBS optical signal collected by the LIBS module needs to accurately represent the composition of the ore sample. The inventors have discovered that two specific factors may need to be controlled to achieve sufficient accuracy in real-time sorting of ore materials outside the laboratory: First, the size of the analytical spot, i.e., the diameter of the LIBS laser beam at its intersection with the top surface of the ore sample, should be small enough that the ion species in the plasma represent only one or a few mineral components of the ore sample. Second, the optical path should be substantially free of dust aerosols or other particles that could contaminate the collected LIBS optical signal.

[0061] Focused control

[0062] Refer again Figure 1 According to one aspect, system 20 includes components such as the “autofocus module” 60 referred to herein, which are configured to adjust the focus of LIBS laser beam 32 in real time to move the analysis spot 36 perpendicular to the conveyor 22 according to the height of the ore sample 24.

[0063] As mentioned above, a laser spot size within the 100 μm range is preferred for obtaining good mineralogical LIBS data. Achieving a consistent spot size on a moving conveyor with varying rock sizes, while maintaining synchronization with the laser repetition rate, is challenging. The correlation between the distance to the ore surface and the measurement must be accurate during laser pulse emission, but not necessarily calculated or accurate outside the emission and detection intervals. In some embodiments, this challenge is addressed by moving the focusing lens 25, which adjusts the focus of the LIBS laser beam to provide a constant spot size on the sample surface in a timely manner, taking into account the passage of the ore sample. This causes the focal plane of the LIBS laser beam to intersect with the top surface of the ore sample passing through the optical path.

[0064] Figure 3A and 3B This is a schematic block diagram of an embodiment of the autofocus module 60. In some embodiments, the autofocus module 60 typically includes a height measuring device, such as a distance sensor 62, configured to measure in real time the sample height of the top surface of the ore sample stream at a point upstream of the LIBS module 30 and along the conveyor travel axis X intersecting the LIBS laser beam 32. The autofocus module 60 also includes a conveyor speed measuring mechanism 64 and a translational lens holder 70 for vertically shifting the focusing lens 25. The conveyor speed measuring mechanism 64 is configured to provide real-time measurement of the travel speed of the ore sample on the conveyor 22. The autofocus module also includes a focus controller 72 that communicates with the height measuring device, the conveyor speed measuring mechanism, the translational lens holder, and the LIBS module, and is configured to focus the LIBS laser beam at the sample height in synchronization with the travel of the ore sample and the light pulses. The focusing controller 72 preferably operates in real time, taking as input a transport speed signal 74 measured by the transport speed measuring mechanism 64; a variable sample height signal 76 measured by the distance sensor 62; and a laser repetition rate 80 from the LIBS module 30, and outputting a focusing control signal 82 to the lens holder 70 to operate the rapidly moving focus. The focusing controller 72 may include an FPGA 75 to provide a precise clock signal, thereby controlling the timing of the focusing control signal with desired real-time accuracy.

[0065] In the illustrated embodiment, distance sensor 62 is located above conveyor 22, upstream of LIBS module 30. Distance sensor 62 is configured to measure the height or vertical position of ore sample 24 as it is conveyed along conveyor path 26. For example, the distance sensor can be implemented using a laser rangefinder having a spot size comparable to (i.e., of the same order of magnitude, preferably between 1 / 3 and 3 times, more preferably between 2 / 3 and 3 / 2) the spot size of LIBS laser beam 32.

[0066] The conveying speed measuring mechanism 64 can be any device, combination of devices, or system that provides real-time measurement of the speed of the conveyor 22.

[0067] refer to Figure 3AIn some embodiments, the conveyor speed measuring mechanism 64 includes a rotary encoder 66 in contact with a portion of the conveyor 22. The encoder rotates at a speed matching the travel speed of the ore sample 24 on the conveyor 22, which can be, for example, a conveyor belt or a turntable on which a conveyor belt is mounted. Therefore, the rotary encoder 66 rotates at the same speed as the ore sample 24. The speed of the conveyor 22 can be calculated using the known number of encoder pulses per rotation and the known encoder wheel diameter, by measuring the frequency of the pulses. This method works well, reliably, and responds well to rapid speed changes, but a drawback is that it requires contact with a moving surface.

[0068] refer to Figure 3B Alternatively, the conveyor speed measuring mechanism 64 can employ a non-contact optical speed measurement method using cross-correlation. This method may involve using a first distance sensor and a second distance sensor located above the conveyor, on the same plane parallel to the surface of the conveyor 22, separated by a predetermined spacing along the conveyor's travel axis. In the illustrated embodiment, a distance sensor 62 located upstream of the LIBS module 30 defines the first distance sensor, and a second distance sensor 68 may preferably be located downstream of the first distance sensor 62, along the conveyor's travel axis. Figure 3B The second distance sensor 68 is precisely aligned with the first distance sensor 62 along the X-axis of the reference frame. In some variations, the second distance sensor 68 is aligned collinearly with the LIBS laser beam 32 in the vertical direction to facilitate alignment testing. In other variations, the second distance sensor 68 can be positioned at any other location that allows for the calculation of the conveyor speed between the two sensors. The second sensor data from the second distance sensor 68 can be provided to the control electronics 73 of the focusing controller via the second sample height signal 77 for processing.

[0069] In some implementations, distance profiles from two distance sensors for N acquisition points are stored at a slower sampling rate, and cross-correlation is performed to extract the time delay between the two profiles. Assuming the alignment between the two sensors is sufficiently precise, the cross-correlation peak will precisely correspond to the time delay between the two distance profiles. Knowing the precise distance between the two distance sensors, the velocity can be calculated using the following relationship:

[0070]

[0071] Where S is the speed of the conveyor, d is the distance between the two distance sensors, and t is the time delay between the two distance profiles.

[0072] For example, the inventors achieved good results in laboratory experiments implementing a velocity measurement technique using two distance sensors at a distance of 5 cm, with a sampling rate of 500 Hz–2 kHz and 512–2048 sampling points. Stable values ​​were provided for at least a 10-fold rolling average. This method may require a minimum speed to work well without requiring a large amount of memory. Cross-correlation is performed on the same FPGA responsible for the fast focusing system to minimize latency in focus calculations using this speed. Other experimental details include: the FPGA used is a De0-nano development board (Altera EP4CE22F17); the main FPGA clock runs at 50 MHz with good results, but it could run even faster using an internal PLL; the laser repetition rate is 100 Hz; and the XY2-100 protocol is used as the digital protocol to drive the focusing lens. The XY2-100 protocol is a 20-bit protocol operating between 2MHz and 4MHz; the ADC is an 8-channel 12-bit multiplexed ADC with a maximum operating speed of 200ksps (ADC128S022); supporting electronics for some signal conditioning of the distance sensor are mounted on the prototype board before the ADC to reduce the voltage and convert the 4mA-20mA current signal into a usable voltage; a differential driver is used to drive the XY2-100 protocol; additional DC-DC converters and circuitry for powering the triggers, encoders, and all supporting peripherals are added; two different models of distance sensors were tested: a commercial OD-150 from SICK and a homemade OD-150 using a PSM-10PSD sensor with an OT301SL transimpedance amplifier from ON-TRAK.

[0073] The focusing controller 72 can be implemented by any device, circuit, system, or combination thereof that provides the required functionality. In the illustrated embodiment, the focusing controller 72 includes control electronics 73 that receives a variable sample height signal 76, a conveyor speed signal 74 or a second sample height signal, and a laser repetition rate 80, and outputs a focusing control signal 82 to the lens holder 70 in precise synchronization with the laser pulses of the LIBS laser beam 32. In the illustrated embodiment, the FPGA 75 is responsible for creating a trigger signal for the laser controller or using the trigger signal as an internal reference clock.

[0074] In one example of the implementation, the autofocus module 60 operates as follows: precisely adjusting and measuring the distance (along the X-axis) and alignment (along the Y-axis) between the distance sensor 62 and the LIBS module 30. For example, an FPGA 75 equipped with an analog-to-digital converter (ADC) samples the variable sampling height signal 76 from the distance sensor 62 at a sampling frequency (10 kHz) higher than the laser repetition rate (100 Hz) of the LIBS laser beam. At each sampling of the Z-axis measurement, the conveyor speed obtained from the analysis of the conveyor speed signal 74 or the variable sample height signal 76 and the second sample height signal 77, along with the X-axis distance constant, is used to calculate the remaining time before the ore sample 24 will reach the LIBS laser position. This remaining time is then compared with all future laser Q-switch triggers, and once a timing match is found, the focusing lens position is calculated. The lens position is calculated using the distance measurement from the distance sensor 62 and a pre-calibrated curve of the distance of the focusing lens 44 relative to its position. This curve can be calibrated once during initial setup and can be updated via software at any future time to push new values ​​to the FPGA 75.

[0075] The FPGA 75 then stores the focusing lens position value in memory and sends it to the lens holder 70 at the correct time to ensure the laser spot is focused. To ensure the focus is stable during the laser pulse, and because the position is known in advance due to the stored value, in some variations, the move command is sent with a predetermined lead delay (e.g., 9 ms) before the laser pulse. This provides time for the focusing lens to stabilize after a rapid movement before the laser pulse arrives. After laser emission, the focusing lens will have a stabilization (no movement) window of at least 1 ms before moving to the next position, during which it collects the LIBS light. These delay values ​​can also be adjusted in software.

[0076] Multiple laser pulses are likely to occur before the ore sample moving below the distance sensor 62 reaches the laser position at different conveyor speeds. To address this, a dynamic FIFO (First-In, First-Out) table can be used to pre-store future lens position values ​​and acquire the correct position at the correct time. This FIFO can automatically adjust its size to accommodate changes in the conveyor speed for future samples.

[0077] airflow system

[0078] Mineralogical analysis of ore samples using LIBS measurements relies on interpreting the relative proportions of the elemental composition of a given mineral to the signal received by the spectrally resolved photodetector. The presence of dust or other particles in the laser beam path introduces noise that is detrimental to mineralogical analysis. In fact, as the focused short laser pulse of the LIBS laser beam propagates toward the rock surface, the light flux increases and can become very high, for example, hundreds of J / cm². 2 Any airborne dust particles or aerosols passing through the optical path of the LIBS laser beam in this region can be absorbed and generate unwanted plasma, adding noise to the signal reaching the spectrally resolved photodetector. Airborne dust may originate from the ambient air of the system's operating environment or be caused by the transition of jet material generated during the LIBS ablation process (e.g., particles generated by shock waves). The latter phenomenon increases with the laser repetition rate because the repetitive shock waves generated by the plasma propel more particles back to the laser source.

[0079] refer to Figure 4 In some embodiments, system 20 includes airflow system 90 configured to generate at least one airflow that moves particles away from optical path 35.

[0080] refer to Figure 4 In some embodiments, the airflow system 90 includes a main nozzle 92 mounted between the LIBS module 30 and the delivery path 26. The main nozzle 92 is preferably formed as a frustoconical shape, tapering gradually from its upper end 94 to its lower end 96. Both ends allow the propagation of both the LIBS laser beam 32 and the plasma light 42. The main nozzle 92 guides an escort airflow 98 that keeps the optical path 35 free of airborne dust and other particles. The airflow system 90 may include a main blower unit 91 that generates the escort airflow 98 and is connected to the main nozzle 92 (e.g., via one or more hoses or flexible conduits) to inject the escort airflow 98 into the main nozzle 92 near the upper end 94. In some embodiments, the upper end 94 of the main nozzle 92 is closed by a top wall 95 to prevent airflow but allow light to pass through. For example, the top wall 95 may be implemented by an optical window made of a material that is light-transmitting at least at the wavelengths of the LIBS laser beam and the plasma light, such as a sheet or film of glass, plastic, etc. The opening 97 is configured to pass through the sidewall of the main nozzle 92 near the upper end 94 to allow the escort airflow 98 to enter, flow downward in the main nozzle, and exit at the open lower end 96. Thus, the escort airflow 98 removes dust and other particles from the optical path 35, "escorting" these particles downward into the conveying path.

[0081] In some embodiments, the airflow system 90 further includes an auxiliary nozzle 100 adjacent to the main nozzle 92 and oriented at a small angle to the optical path 35. The auxiliary nozzle 100 generates a clean airflow 102 toward the general area where the optical path 35 of the LIBS system intersects with the delivery path 26. The clean airflow 102 is preferably configured to "push away" particles in the analytical spot region, including particles output from the main nozzle, particles generated by the shock wave, and ambient air particles, leaving a substantially particle-free optical path 35.

[0082] The airflow system 90 may include an auxiliary blower or air compressor unit 101 that generates a cleaning airflow 102 and is connected to the auxiliary nozzle 100, for example, via one or more hoses or flexible conduits. An opening 103 is provided through the sidewall of the auxiliary nozzle 100 to allow the input of the cleaning airflow 102. Since the shape of the auxiliary nozzle 100 is not limited by the optical path 35, the cleaning airflow 102 can include a variety of airflow velocities and shapes, using less air than the main nozzle 92. Preferably, the escort airflow 98 from the main nozzle 92 and the cleaning airflow 102 from the auxiliary nozzle 100 generate an airflow that allows coverage of multiple sample height values ​​without adjustment: the main nozzle 92 is collinear with the optical path 35, and the cleaning airflow 102 from the auxiliary nozzle 100 is shaped to cover the entire range of sample heights. It should be noted that since the plasma is generated in a very short time on the order of nanoseconds, and the LIBS measurement is performed within microseconds, the timescale of the LIBS measurement is much shorter than the timescale of the purging. Therefore, the purging process does not have time to remove the plasma before the plasma light is generated.

[0083] In some embodiments, the airflow system 90 may include a scraper nozzle 104 positioned upstream of the LIBS module 30 and above the transport path 26. The scraper nozzle 104 is configured and shaped to generate a sufficiently strong scraper airflow 106 to remove unwanted material from the surface 23 of the ore sample 24. Advantageously, the scraper airflow 104 can remove deposited material, such as mud or dust, from the top surface of the ore sample, which may obstruct the path of the LIBS laser beam. The airflow system 90 may include a scraper blower or air compressor unit 107 that generates the scraper airflow 106 and is connected to the scraper nozzle 104 through a sidewall opening 105.

[0084] refer to Figure 5A and 5BIn some embodiments, system 20 includes one or more protective mechanisms to prevent ore samples on the transport path from damaging components of the airflow system. The rock flow on the transport path may have a profile exceeding the useful optical range. Since air nozzles 92, 100, and 104 are the lowest parts of the system, such rock flow could impact the nozzles and cause damage.

[0085] In one embodiment, the protective mechanism includes a protective beam 108 or plowshare 108' positioned upstream of a first nozzle in the nozzle (and even upstream of the first distance sensor 62) along the conveying path 26. In some embodiments, when impacted by the protective beam 108, an excessively large rock 24 is pushed back and / or rolled, or, if using the plowshare 108', is deflected. The protective beam 108 can be a robust beam of any shape, transverse to the conveying path 26 and thus transverse to the movement of the ore sample 24, or, in the case of the plowshare 108, positioned at an angle. In a typical embodiment, the protective beam can be a box beam or h-beam of steel, but any material capable of withstanding the impact of an excessively large rock can be used. The same function can be achieved by a plowshare with a similar geometry or blade geometry.

[0086] In another embodiment, the main nozzle, auxiliary nozzle, and / or scraper nozzle may be attached to the hinge system 110. In some variations, the LIBS module 30 may be in a low position to allow for the detection of a maximum number of samples, while the hinge system prevents final nozzle-rock impact damage to the system. In another variation, nozzles made of flexible materials may be used.

[0087] Processing unit

[0088] Refer again Figure 1 The system 20 also includes a processing unit 120 for performing mineralogical analysis of ore samples based on LIBS optical signals. The processing unit 120 may include a processor 124 and a memory 126.

[0089] Processor 124 can run an operating system and execute computer programs, also known as commands, instructions, functions, procedures, software code, executable files, application programs, etc. (Although for illustrative purposes...) Figure 1The processor 124 is depicted as a single entity, but the term "processor" should not be construed as limited to a single processor; therefore, any known processor architecture may be used. In some embodiments, the processor 124 may include multiple processing units. Such processing units may be physically located within the same device, or the processor 124 may represent the processing capabilities of multiple devices operating in cooperation. For example, the processor 124 may include, or may be a subset of, one or more of the following: a computer; a microprocessor; a microcontroller; a coprocessor; a central processing unit (CPU); an image signal processor (ISP); a digital signal processor (DSP) running on a system-on-a-chip (SoC); a single-board computer (SBC); a dedicated graphics processing unit (GPU); a dedicated programmable logic device implemented in a hardware device, such as, for example, a field-programmable gate array (FPGA) or an application-specific integrated circuit (ASIC); a digital processor; an analog processor; digital circuits designed for processing information; analog circuits designed for processing information; a state machine; and / or other mechanisms configured to electronically process information and operate jointly as processors.

[0090] The memory 126, also referred to as a "computer-readable storage medium," is capable of storing computer programs and other data to be retrieved by the processor 124. The stored data may include LIBS optical signals and information extracted therefrom. The terms "computer-readable storage medium" and "computer-readable memory" herein refer to a non-transitory and tangible computer product capable of storing and transmitting executable instructions for performing the various steps of the techniques disclosed herein. Computer-readable memory can be any computer data storage device or component of such a device, including random access memory (RAM); dynamic RAM; read-only memory (ROM); magnetic storage devices such as hard disk drives, solid-state drives, floppy disks, and magnetic tapes; and optical storage devices such as optical discs (CDs or CD-ROMs), digital video discs (DVDs), and Blu-ray discs. TM Optical discs; flash memory; and / or any other non-transitory memory technology. As will be understood by those skilled in the art, multiple such storage devices may be provided. The computer-readable storage device may be associated with, coupled to, or included in a processor configured to execute instructions contained in a computer program stored in the computer-readable storage device and relating to various functions associated with the processor. Furthermore, the processor 124 and the memory 126 of its components may be part of or separate from the LIBS module 30.

[0091] Since LIBS generates elemental spectra, the goal is to convert these spectra into mineralogical features. As mentioned above, one of the features and systems described herein is to control the size of the analytical spots so that they are comparable to the size of one or more mineral components. Larger spot sizes used for LIBS measurements result in rather complex spectra of different minerals present in the ablation spots. In some implementations, advanced chemometric data processing methods known in the art for deconvolution of mixed spectra can be used to identify and quantify individual mineral features in the mixture. A description of an example of a mineral analysis method for laboratory proof-of-concept can be found in El Haddad et al., “Multiphase mineral identification and quantification by laser-induced breakdown spectroscopy” (Minerals Engineering, (2019), Vol. 134, pp. 281-290), the entire contents of which are incorporated herein by reference.

[0092] Example

[0093] A set of rock tiles from an Australian porphyry copper deposit were characterized using QMA, and the data were used to guide and validate results obtained via LIBS. LIBS maps were obtained using a step size of 100 μm and a spot size of 70 μm. To compare QMA and LIBS maps, the spatial resolution of the QMA map had to be reduced by an average of 400 points. Unsupervised multivariate curve resolution-alternating least squares (MCR-ALS) provided optimal results for the identification, quantification, and imaging of minerals on rock tiles, even in the presence of mixed mineral phases within the laser spot regions (El Haddad et al., 2019).

[0094] The mineral abundance and imaging of the selected mineral phases in this work were successfully obtained, including bornite, chalcopyrite, pyrite, molybdenite, quartz, chlorite, potassium feldspar, sodium feldspar, fluorite, and calcite. As mentioned earlier, one of the main challenges of this method is deconvolution of mixed spectra. Limiting the size of the analytical spot to the size of the mineral composition limits the number of different minerals in the analysis; for example, an average of 2 or 3 different minerals are present in a 70 μm probe spot, making the challenge more manageable. The method was first calibrated using LIBS and QMA data acquired at the same location. The model was validated by processing a test set of LIBS data from different locations on the tile and comparing the results with QMA data. The test set contained 39 20×20 pixel images. Figure 6A and 6B This is the result of LIBs ( Figure 6A ) and QMA quantitative results ( Figure 6A The mineral abundance images are compared. In both cases, the pixel color represents the mineral with the highest abundance at a given spatial location, and the color gradient represents its abundance value. It should be noted that the abundance values ​​of all minerals were measured by LIBS at each spatial location. The mineral distribution measured by LIBS is in excellent agreement with the corresponding QMA data used as a reference. The predicted mineral abundance measured by each LIBS spot is in excellent agreement with the reference QMA value, confirming the quantitative capability of the model.

[0095] Similarly, in the point counting method, the abundance of a given mineral phase is determined by the ratio of the number of measurement points where the mineral phase was detected to the total number of measurement points. For LIBS measurements, mineral quantification is also performed in each measurement because larger spot sizes typically cover more than one mineral phase, while phase abundance is determined in a single measurement. The results of mineral phase abundance predictions across the entire map are shown in Figure 7. The deviation of the values ​​obtained from LIBS and QMA is specified as absolute error. For the 10 mineral phases analyzed, this difference is less than 4%.

[0096] Following encouraging results obtained on porphyry copper ceramic tiles in spectral mode, and more importantly in point-count mode, the next step is to evaluate the robustness of both the LIBS technique and the chemometric data processing method in more realistic settings. Several studies were conducted to assess the method's sensitivity to laser energy and spot size, examine the effects of unpolished rock surfaces, the presence of dust, and applicable LIBS sampling strategies. Some preliminary results and conclusions from these studies are presented below.

[0097] There is no obvious reason why rock surface roughness alters the LIBS emission signal. However, to avoid any unforeseen consequences later, the robustness of the prediction must be tested on unpolished tiles or rocks. A series of tests were performed on single-mineral rocks using the system described in this paper. The mineral rocks from the porphyry copper deposit are from Ward's... Purchased. The model was tested on both polished and unpolished surfaces and simulated the surface angle distribution at different angles. A second method measured a sandblasted single-mineral rock surface and compared the predicted measurements to polished tiles. A third method involved using a third LIBS laser blast at the same location to simulate a natural surface.

[0098] The predictive power of the method was validated by probing quartz, microcline, and calcite rocks using LIBS. Quartz was identified as 99% quartz, but surprisingly, microcline (potassium feldspar) was identified as 77% potassium feldspar and 10% sodium feldspar, while calcite was identified as 71% calcite and 26% quartz. The rocks were then characterized by X-ray diffraction (XRD), confirming the presence of other minerals and the model predictions.

[0099] LIBS measurements were performed at different angles on a single mineral rock to simulate different facet angles on a real rock surface. The surface was also roughened with sandpaper for 50 seconds and 500 seconds to simulate a real, unpolished rock surface. For each case, measurements were taken at 10 different locations on the sample, repeated 3 times at each location. The spectra from the 3 replicates were then averaged and processed by the model. The results were in good agreement with those obtained under normal incidence. Ward's quartz results showed very stable predictions from 98% to 100% quartz under various conditions, as expected. Ward's calcite results showed greater fluctuations in predictions for both calcite and quartz, likely due to the real variation in quartz to calcite from one location to another. Ward's microcline results also showed fluctuations, which could also be explained by real mineral concentration variations, but in addition, there were fluctuations in different predictions for unknown or unidentified minerals. These fluctuations are attributed to the model's performance in correctly deconvolving the characteristics of minerals composed of the same elements in relatively similar proportions. For example, albite (NaAlSi3O8) and potassium feldspar (KAlSi3O8) are separated solely by the presence of Na and K. If any of these emission lines is saturated, the quantitative analysis decreases. In all cases tested, no significant correlation was found between different conditions (angle, roughness, etc.) and predicted fluctuations.

[0100] In the third experiment, three LIBS measurements were performed at each location. The second and third laser beams were incident on the ablation site of the first beam, which simulates a real, unpolished rock surface. The results showed no significant difference or degradation in the predictions for the first, second, and third laser beams. It should be noted that these measurements were performed on porphyry copper tiles, not on a single mineral sample from Ward. Furthermore, the number of data points is quite significant, and the statistics are more representative. The root mean square error (RMSE) values ​​for the first, second, and third laser beams were similar, and R² was greater than 0.9 for all predicted minerals.

[0101] Of course, many modifications can be made to the above embodiments without departing from the scope of protection.

Claims

1. A system for online sorting of ore samples based on mineralogical analysis, comprising: - A conveyor for moving a stream of ore samples along a conveying path, the ore samples having a variable sample height on the conveyor; - LIBS module, which projects a pulsed LIBS laser beam along the optical path, and focuses the pulsed LIBS laser beam onto the analysis spot on the transport path. The size of the analysis spot is on the order of magnitude of the size of one or more mineral components in the ore sample. The LIBS module collects the LIBS light signal returning along the optical path. - A height measuring device configured to measure in real time the sample height of the top surface of the ore sample stream along a conveyor travel axis intersecting the LIBS laser beam at a point upstream of the LIBS module. - A conveyor speed measuring mechanism, configured to provide real-time measurement of the travel speed of the ore sample on the conveyor; - A focusing controller, which communicates with the height measuring device and the conveyor speed measuring mechanism, and is configured to focus the LIBS laser beam at the sample height in sync with the movement of the ore sample and the optical pulse, based on the real-time measurement of the sample height by the height measuring device and the real-time measurement of the travel speed of the ore sample by the conveyor speed measuring mechanism. - An airflow system configured to generate at least one airflow that disperses particles away from the optical path; and - A processing unit, which performs mineralogical analysis on the ore sample based on the LIBS optical signal.

2. The system of claim 1, wherein the LIBS laser beam has a diameter between 70 μm and 140 μm at the analysis spot.

3. The system of claim 1, wherein the LIBS laser beam has a diameter of approximately 100 μm at the analysis spot.

4. The system of claim 1, wherein the LIBS module includes a focusing lens and a translational lens holder, the focusing lens focusing the LIBS laser beam at the analysis spot, and the translational lens holder being configured to move the focusing lens vertically under the control of the focusing controller.

5. The system of claim 1, wherein the conveyor speed measuring mechanism comprises a rotary encoder in contact with a portion of the conveyor, the rotary encoder rotating at a rotational speed matching the travel speed of the ore sample on the conveyor.

6. The system of claim 1, wherein the conveyor speed measuring mechanism includes a first distance sensor and a second distance sensor, the first distance sensor and the second distance sensor being located above the conveyor, on the same plane parallel to the surface of the conveyor, the first distance sensor and the second distance sensor being separated by a predetermined spacing along the conveyor travel axis.

7. The system of claim 6, wherein the second distance sensor is collinearly aligned with the LIBS laser beam in the vertical direction.

8. The system of claim 1, wherein the focusing controller comprises an FPGA configured to sample a variable sample height signal from the height measuring device at a sampling frequency higher than the laser repetition rate of the LIBS laser beam.

9. The system of claim 8, wherein the FPGA includes an analog-to-digital converter.

10. The system of claim 1, wherein the airflow system includes a main nozzle installed between the LIBS module and the delivery path.

11. The system of claim 10, wherein the main nozzle has an upper end and a lower end, the upper end and the lower end allowing the LIBS laser beam and plasma light to propagate through, the main nozzle being shaped as a truncated cone that tapers gradually from the upper end to the lower end.

12. The system of claim 11, wherein the airflow system includes a main blower unit that generates a convective airflow and is connected to the main nozzle to inject the convective airflow into the main nozzle near the upper end.

13. The system of claim 11, wherein the upper end of the main nozzle is closed by a top wall to prevent airflow but allow light to pass through.

14. The system of claim 10, wherein the airflow system includes an auxiliary nozzle adjacent to the main nozzle and oriented at a small angle to the optical path, the auxiliary nozzle generating a clean airflow toward the region where the optical path intersects the delivery path.

15. The system of claim 14, wherein the airflow system includes an auxiliary blower unit that generates the clean airflow and is connected to the auxiliary nozzle.

16. The system according to any one of claims 1 to 15, wherein the airflow system includes a scraper nozzle disposed upstream of the LIBS module and above the transport path, and the scraper nozzle is configured and shaped to generate a sufficiently strong scraper airflow to remove unwanted material from the surface of the ore sample.

17. The system according to any one of claims 1 to 15, the system comprising one or more protective mechanisms that prevent ore samples on the transport path from damaging components of the airflow system.

18. The system according to any one of claims 1 to 15, wherein the mineralogical analysis performed by the processing unit includes identifying and quantifying individual mineral features of the composition of the ore sample using chemometric data processing methods for mixed-spectral deconvolution.

Images