Method for detecting indirect indicators of mill lining performance and wear using IoT sensors

Abstract

The present invention proposes a method for monitoring wear in ball mills, using wireless sensors and advanced data processing techniques. Wireless sensors are strategically positioned on the mill lining and collect vibration data, which is then processed to identify characteristic patterns that indicate both grinding quality and lining wear. This inventive solution eliminates the need for intrusive and expensive additional devices, reducing downtime and increasing equipment availability. Vibration data analysis is performed in both the time and angle domains, allowing the identification of indirect indicators of lining wear, and also allowing the evaluation of grinding quality. The internet communication interface and integration with cloud services provide operators with real-time access to mill operation, facilitating continuous monitoring and proactive decision-making for process maintenance and optimization.

Description

[0001] METHOD FOR DETECTING INDIRECT INDICATORS OF MILL LINING PERFORMANCE AND WEAR USING IOT SENSORS

[0002] FIELD OF THE INVENTION

[0003]

[0001] The present invention relates to the technical field of minerals processing and comminution, particularly to indicators for evaluating the wear of mill linings. The efficiency and reliability of this equipment depend significantly on the health of the internal linings, which undergo considerable wear and degradation during operation. The present invention presents a technical solution for detecting indirect indicators of mill lining performance and wear, thus facilitating the implementation of proactive maintenance strategies and optimizing the overall operation of the equipment.

[0004] STATE OF THE ART

[0005]

[0002] Ball mills are widely used in comminution processes to reduce the size of raw materials in industrial sectors such as mining, cement, quarrying and fertilizer sectors. The operating principle involves the rotation of a cylinder filled with steel balls, which act on the material to be ground by impact and abrasion. Duringthis process, the mill lining plates, which protect its structure and improve grinding efficiency, undergo continuous wear. Failure to properly monitor this wear can result in equipment failures, unplanned outages, and high maintenance costs.

[0006]

[0003] Usually, the evaluation of lining wear is done through scheduled physical inspections. However, these methods are subjective and have a number of limitations, such as the long interval between inspections, allowing wear to evolve without detection. However, these methods are subjective and have a number of limitations, such as the long interval between inspections, allowing wear to increase without being detected.

[0004] Recent technologies have introduced several methodologies for monitoring the structural and operational health of ball mills. Some of these methods are based on the analysis of acoustic signals generated inside the mill, or on the movement of grinding balls during operation, as described in document KR101620507B1. However, these solutions have limitations, as they may not provide a complete assessment of the mill's health, and do not offer a comprehensive view of its operating conditions.

[0007]

[0005] Other techniques include installing multiple sensors at locations within the mill, as disclosed in document CN218502230U. Although they provide direct data on wear, these solutions are often complex and invasive, requiring regular maintenance and frequent replacement of sensors, which increases operational costs and interrupts the production process.

[0008]

[0006] In addition, there are methodologies that rely heavily on a combination of multiple operational data sources, such as engine power or mill load and filling levels, as described in document US2017225172A1. While such solutions offer a multifactorial view, the need to integrate different variables can introduce additional complexity to the monitoring process, and also create dependence on operational data that, by their nature, can vary according to external conditions, reducing the accuracy of wear diagnosis.

[0009]

[0007] Therefore, there is a need for a simplified and efficient solution for monitoring the health of ball mills, thus reducing dependence on additional equipment and multiple sources of data and correlations, enabling the simplification of the monitoring process and improved operational autonomy. A promising option to meet these requirements is the use of vibration data processing.

[0010] OBJECT OF THE INVENTION

[0008] The object of the present invention is based on the need to accurately monitor the wear of ball mill linings, a critical component in the operation of comminution processes. The innovative solution consists of installing wireless vibration sensors, positioned externally to the mill, which collect data in real time during equipment operation.

[0011]

[0009] This configuration eliminates the invasive complexity associated with the installation of internal sensors, as disclosed in document CN218502230U, which requires multiple sensors installed in specific locations within the mill and which often require regular maintenance and replacement, causing increased operating costs and interruption of the production process.

[0012]

[0010] Furthermore, the invention distinguishes itself from other existing solutions, such as those that use optical signals for monitoring. Wireless data transmission technology ensures that information is sent to a central processing unit, where advanced analytics algorithms can identify patterns and predict maintenance needs.

[0013] [Oil] In this way, the invention not only optimizes monitoring efficiency, but also contributes to reducing operating costs and maximizing the operating life of equipment, becoming an essential solution for industries that depend on grinding materials. This combination of factors highlights the novelty and inventive step of the present invention, which offers an effective response to the limitations of existing technologies.

[0014] BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 represents the diagram of a wireless collection monitoring system for vibration detection and analysis in a ball mill.

[0016] Figure 2 represents the flowchart of steps for acquiring the indicators. Figure 3 represents the indicators in polar coordinates.

[0017] Figure 4 represents the trend of indicators as a function of time. DESCRIPTION OF THE INVENTION

[0018]

[0012] The present invention relates to the area of condition monitoring system (10) of ball mills (11), as illustrated in Figure 1. The ball mill (11) operates by means of a drive system (20), for example by means of an electric motor or other type of device, which causes the cylinder (21) to rotate around its longitudinal axis. Inside the cylinder (21), grinding bodies (22) are arranged, which can be balls made of steel, ceramic or other wear-resistant materials. As the mill body rotates, these bodies (22) move, come into friction and collide with each other and with the material to be ground, promoting fragmentation and reducing the size of the particles. Usually, monitoring the wear of the grinding bodies (22) and lining (13) of assets is performed using methods such as manual ultrasound analysis and visual inspections during asset shutdown. However, these methods have limitations such as the need for frequent shutdowns, high frequency and subjectivity in the analysis.

[0019]

[0013] The core of the present method is the strategic implementation of at least one wireless vibration sensor (12) in the mill structure (11) to collect vibration signals from the asset structure during its operation, providing the foundation for predicting operational events and early identification of problems in the grinding process.

[0020]

[0014] The present invention proposes a solution for monitoring wear in ball mills (11), using wireless sensors (12) and advanced data processing techniques. The wireless vibration sensors (12) are strategically positioned on the mill body (11), between the connecting rings or flanges, and their physical connection can be made by means of glue, screws or both on the external surface of the ball mill.

[0021]

[0015] The captured signals are processed to identify characteristic patterns that indicate both grinding quality and lining wear (13). Wireless communication with each sensor unit (12) is initiated via a collection agent (14 or 15). In this communication, the measurement data stored in the memory of the devices are synchronized.

[0022]

[0016] Additionally, wireless sensors use MEMS accelerometers, which, in addition to measuring the vibrational component of the mill (11), can measure the static component of acceleration (acceleration of gravity), which can be used as a rotational speed meter when attached to rotating parts. This aspect is crucial to the present method.

[0023]

[0017] In another aspect of the present method, the position of the sensors (12) refers to the feed, central and discharge region of the mill (11), with the preferred positioning configuration being at least one sensor (12), in each of the rings, so that monitoring these regions separately contributes to knowledge of load distribution. As an alternative configuration, it is possible to perform monitoring by installing more than one sensor (12) distributed around the same ring, as long as there is at least one in each region. The positioning of the sensor (12) in each region must be done so that it is appropriately exposed to the excitations caused by grinding.

[0024]

[0018] By integrating a wireless data transmission module and a gateway (15) or mobile (14), the collected data is securely transmitted to a remote processing unit, enabling advanced analysis in real time. Furthermore, through an internet communication interface (16), vibration signals are accessible on mobile devices (14), allowing operators to view the mill's operation in real time.

[0025]

[0019] Integration with cloud services (17) enables secure data storage and access to online software (19) for more detailed analysis. Through a data request interface (18), users can access calculated metrics and receive alerts if acceptable operating limits are exceeded or predetermined wear is detected. In summary, the present invention includes integration with modern or Internet of Things (loT) technologies, such as wireless data transmission devices (14 or 15), Internet communication (16) and cloud storage (17), providing remote and realtime access to data through a central platform (19).

[0026]

[0020] The present invention is advantageous in relation to the state of the art, overcoming the limitations and challenges faced by conventional means of condition monitoring (10) of the ball mill (11). While conventional approaches rely heavily on scheduled physical inspections and technical expertise, the present invention takes a proactive, data-driven approach, thus minimizing the subjectivity and time required to detect potential problems. In contrast to other methodologies that can be intrusive or complex, requiring the installation of multiple sensors in specific internal locations of the mill, the proposed monitoring structure is lean and simple.

[0027]

[0021] To calculate the instantaneous position of the sensor (12) during the rotation of the cylinder (21), the acceleration of gravity is used, and for this purpose the raw signal (23), as illustrated in FIG. 2, acquired on the three axes (axial, tangential and radial) needs to be divided into low (24) and high frequency (25) by means of dedicated filters. This signal segregation allows the independent examination of high frequency vibrations (25), generally associated with impacts and localized events, and of low frequency components (24) that allow the detection of the maximum points of the radial and / or tangential axis and use this point as a reference for the start of rotation.

[0028]

[0022] The flowchart (26) in FIG. 3 illustrates in detail the steps for analyzing and processing vibration data from the mill (11) in order to extract relevant metrics for condition monitoring and predictive maintenance of internal linings (13).

[0029]

[0023] The process begins with the acquisition of the signal (27), which captures the raw vibration signal (23). From this data, two analysis approaches are followed simultaneously. The first approach applies a demodulation technique to extract the envelope (28), from which the metrics related to the signal impulsivity (envelope metrics) (29) are calculated, important for identifying the loss of stiffness of the lining (13).

[0030]

[0024] In parallel, the second approach separates the raw signal (30) into low and high frequency components. The low frequency component (24) is used to calculate the phase of the signal (31), which represents the angular position of the mill shaft at each instant. Phase is critical in determining the size of analysis intervals (32), ensuring that each window captures a region of adequate size for accurate detection of critical events.

[0031]

[0025] The resulting signal is then segmented by revolution (35), characterizing the vibration behavior of each mill rotation cycle separately. In an additional aspect of the proposed segmentation, it is possible to determine the exact velocity value of the cylindrical body (21) (33), which can be used as an indicator of movement performance.

[0032]

[0026] The resulting acceleration signals (37) and the resulting acceleration envelope signals (36) are extracted to assist in the process of visually characterizing the performance angles, thus enabling alternative representations of the vibration that can assist in identifying patterns and anomalies.

[0033]

[0027] The fall angles (40) and lifting angles (41) are calculated (38) from the expected vibration energy of each grinding region, representing the regions of impact and detachment of the grinding body, respectively. The grinding angle (43) is obtained from the difference between the two aforementioned angles, thus providing information on the efficiency of the grinding process.

[0034]

[0028] In addition to the angles, the fall (39) and lifting (42) energies are extracted, which quantify the vibration intensity at these specific moments of the grinding process. Additionally, the polar windowed signal (37) and the polar envelope (36) are generated.

[0035]

[0029] The lifting angles (41) and the fall angles (40) are important for the real understanding of the dynamic operation of the mill. The fall and lifting points that indicate the point of impact of the bodies (22) and their detachment, respectively. These points may indicate the "cascade" (desired) or "cataract" (undesired) grinding format, the distribution of the material and / or grinding body in the mill (11) and the wear of the lining itself (13). Furthermore, when these indicators are compared in sections of the mill (11) (feed, central and discharge region) they highlight some problems such as inappropriate feeding or accumulation of material.

[0036]

[0030] By applying the method, a large amount of data is obtained, which allows operators and maintenance teams to have valuable information about the health of the ball mill (11) and its potential failures. One of the most powerful tools for understanding its performance is the polar graph (44).

[0037]

[0031] In this aspect, the polar graph (44), presented in FIG. 4, facilitates the simultaneous evaluation of several metrics extracted through the proposed process, which includes the lifting angle (41), and functions as a fingerprint of the point of impact between the grinding bodies (22) and the lining (13) of the mill (11). The fall angle (40) indicates the beginning of a region with minimum vibration energy, the grinding angle (43), calculated from these two angles, offers an additional perspective on the grinding dynamics.

[0038]

[0032] Spectral entropy and instantaneous RMS techniques can be used to quantify the complexity of the vibration signal, potentially reflecting irregularities or wear patterns in the lining (13). The acceleration envelope technique (36) represents the upper limit of the vibration signal, which may indicate extreme impact events.

[0033] Additionally, lifting energy (42) and fall energy (39) are local vibration energy metrics calculated within their respective windows. By analyzing the distribution and these metrics trends on the polargraph, operators gain general insights into the health of the mill, enabling the identification of possible

[0039]

[0034] FIG 4 shows the time-trend graphs of the metrics obtained by processing the signals from the sensors installed along the mill. This way, each monitoring point in the mill (11) will have its own set of these 5 indicators, which facilitates cross-validation and problem diagnosis.

[0040]

[0035] The lifting angles (41) and the fall angles (40) are design parameters of the mill (11) and should operate as close to ideal as possible for efficient grinding. In one example, the decrease or increase in the lifting angle (41) of the balls may be related to the fall or increase in the degree of filling, wear or otherwise of the lining lifters (13) and the decrease or increase in the rotation speed (33), respectively.

[0041]

[0036] In either case, it is interesting to act on the change of this angle (41) to improve the efficiency of the process. The fall angle (40) is inversely proportional to the filling, but presents the same proportional relationship of the lifting angle (41) with the lining wear (13) or mill speed. The grinding angle graph (43) is strongly correlated and proportional to the mill filling (11).

[0042]

[0037] Additionally, monitoring the vibration energy of the lifting angle (41) and fall angle (40) can provide strong evidence of a worn lining (13) (if levels decrease) or an increase or decrease in filling if the energy (39, 42) increases or decreases respectively.

[0043]

[0038] The graph of lifting angle behavior (41) over time demonstrates in a metric that points out the moment of maximum impact between the material being ground and the lining (13). A significant increase in the lifting angle (41) may suggest a failure in the material or a change in the material being processed. On the other hand, a critical decrease in the lifting angle (41) could reflect a breakage of the grinding media (22) or a reduction in the filling level inside the mill.

[0044]

[0039] Thus, trend graphs, when analyzed in conjunction with the polar graph (44), offer a powerful set of diagnostic tools. By understanding the expected behavior of each metric and its correlation to potential failures, operators will be able to use this information for proactive maintenance strategies. Early detection of these problems allows timely intervention, minimizing downtime and ensuring smooth and efficient operation of the ball mill (11).

[0045]

[0040] Therefore, the present invention stands out for its inventive step and novelty, not only for the application of wireless sensors in an industrial context, but mainly for the synergy between data collection, advanced analysis and communication via loT. The present invention overcomes the limitations of traditional condition monitoring methods by allowing early detection of problems such as failures in the lining (13) or in the grinding process, such as the breakage of grinding media (22) or a reduction in the filling level. Therefore, the invention optimizes predictive control, ensuring greater operational efficiency and reducing unexpected downtime.

[0046]

[0041] By understanding the expected behavior of each metric and its correlation to potential failures, operators will be able to use this information for proactive maintenance strategies. Detection allows timely intervention, minimizing downtime and ensuring smooth and efficient operation of the ball mill (11). While traditional methods rely on manual inspections and subjective analysis, the proposed approach uses objective data collected by sensors to predict failures and optimize ball mill performance. Furthermore, the implementation of techniques such as signal demodulation and polar analysis offers a depth of analysis that allows for more effective diagnostics, resulting in a system that not only responds to monitoring challenges but also redefines operational practices in the sector. Thus, the invention presents an innovative and significant contribution to the field of predictive maintenance engineering.

Claims

METHOD FOR DETECTING INDIRECT INDICATORS OF MILL LINING PERFORMANCE AND WEAR USING IOT SENSORSFIELD OF THE INVENTION[001] The present invention relates to the technical field of minerals processing and comminution, particularly to indicators for evaluating the wear of mill linings. The efficiency and reliability of this equipment depend significantly on the health of the internal linings, which undergo considerable wear and degradation during operation. The present invention presents a technical solution for detecting indirect indicators of mill lining performance and wear, thus facilitating the implementation of proactive maintenance strategies and optimizing the overall operation of the equipment.STATE OF THE ART[002] Ball mills are widely used in comminution processes to reduce the size of raw materials in industrial sectors such as mining, cement, quarrying and fertilizer sectors. The operating principle involves the rotation of a cylinder filled with steel balls, which act on the material to be ground by impact and abrasion. Duringthis process, the mill lining plates, which protect its structure and improve grinding efficiency, undergo continuous wear. Failure to properly monitor this wear can result in equipment failures, unplanned outages, and high maintenance costs.[003] Usually, the evaluation of lining wear is done through scheduled physical inspections. However, these methods are subjective and have a number of limitations, such as the long interval between inspections, allowing wear to evolve without detection. However, these methods are subjective and have a number of limitations, such as the long interval between inspections, allowing wear to increase without being detected.[004] Recent technologies have introduced several methodologies for monitoring the structural and operational health of ball mills. Some of these methods are based on the analysis of acoustic signals generated inside the mill, or on the movement of grinding balls during operation, as described in document KR101620507B1. However, these solutions have limitations, as they may not provide a complete assessment of the mill's health, and do not offer a comprehensive view of its operating conditions.[005] Other techniques include installing multiple sensors at locations within the mill, as disclosed in document CN218502230U. Although they provide direct data on wear, these solutions are often complex and invasive, requiring regular maintenance and frequent replacement of sensors, which increases operational costs and interrupts the production process.[006] In addition, there are methodologies that rely heavily on a combination of multiple operational data sources, such as engine power or mill load and filling levels, as described in document US2017225172A1. While such solutions offer a multifactorial view, the need to integrate different variables can introduce additional complexity to the monitoring process, and also create dependence on operational data that, by their nature, can vary according to external conditions, reducing the accuracy of wear diagnosis.[007] Therefore, there is a need for a simplified and efficient solution for monitoring the health of ball mills, thus reducing dependence on additional equipment and multiple sources of data and correlations, enabling the simplification of the monitoring process and improved operational autonomy. A promising option to meet these requirements is the use of vibration data processing.OBJECT OF THE INVENTION

Images