BACKGROUND

Field

[0001] The present disclosure is generally directed to industrial equipment deployed in Internet of Things (loT) systems, and more specifically, to facilitating machine learning based industry equipment working pattern profiling and applications.

Related Art

[0002] In related art implementations, many types of industrial equipment have loT sensors installed, and industrial equipment manufacturers have started to deploy loT platforms to collect equipment operating data in real time. Such loT data can be utilized to build visualization tools on the sensor data directly without advanced analytics, which is easy to implement but not sufficient to provide valuable insights or recommendations for either customers or manufactures.

[0003] The working patterns of industrial equipment such as air compressors affect energy efficiency, equipment health and can also reveal important usage insights that can be used to improve equipment design. For example, for air compressors, if a specific industry sector or process has a certain working pattern such as fast loading from 0 to target discharge pressure level or long time of high pressure output, the manufacturer can ensure the usage pattern to be satisfied during design stage.

[0004] Using machine learning algorithms, unique normal working pattern signatures can be constructed for each piece of equipment that can be used to detect if there is abnormal behavior when new operating data is obtained.

[0005] When performance key performance indicators (KPI) such as energy efficiency are available, mathematical mappings can be built between the working pattern with the KPI using machine learning and hence to improve the KPI. SUMMARY

[0006] The present disclosure is directed to systems and methods to analyze vast amount of industrial equipment loT data collected by manufacturers to build machine learning based working pattern profiling and applications to improve equipment design, equipment heath, energy efficiency, etc. Instead of using ad-hoc profiling based on intuition or domain knowledge, data driven profiling methods are able to identify complex relationships among different sensor values.

[0007] Industrial equipment such as compressors usually have different working modes, such as loading, stably working, offloading, off, and so on. Sensor values will have different patterns under different working modes. It can be more meaningful to extract features based on working modes than based on the related art sliding window or landmark window feature extraction.

[0008] Aspects of the present disclosure involve a computer program, storing instructions for an analytics platform for a system having a plurality of equipment, the instruction involving detecting, from time series sensor data received from the plurality of equipment, starting points across one or more operating cycles including a loading mode of operation, an unloading mode of operation, a stable mode of operation, and an off mode; extracting a plurality of features from the time series sensor data for each of the loading mode of operation, the unloading mode of operation, the stable mode of operation, and the off mode in the one or more operating cycles based on the detected starting points; and generating one or more profiling models from the extracted features for the plurality of equipment. The instructions can be stored in a non- transitory computer readable medium and executed by one or more processors.

[0009] Aspects of the present disclosure involve a method for an analytics platform for a system having a plurality of equipment, the method involving detecting, from time series sensor data received from the plurality of equipment, starting points across one or more operating cycles including a loading mode of operation, an unloading mode of operation, a stable mode of operation, and an off mode; extracting a plurality of features from the time series sensor data for each of the loading mode of operation, the unloading mode of operation, the stable mode of operation, and the off mode in the one or more operating cycles based on the detected starting points; and generating one or more profiling models from the extracted features for the plurality of equipment. [0010] Aspects of the present disclosure involve an analytics platform system for a system involving a plurality of equipment, the system involving means for detecting, from time series sensor data received from the plurality of equipment, starting points across one or more operating cycles including a loading mode of operation, an unloading mode of operation, a stable mode of operation, and an off mode; means for extracting a plurality of features from the time series sensor data for each of the loading mode of operation, the unloading mode of operation, the stable mode of operation, and the off mode in the one or more operating cycles based on the detected starting points; and means for generating one or more profiling models from the extracted features for the plurality of equipment.

[0011] Aspects of the present disclosure involve an apparatus configured to facilitate an analytics platform for a system having a plurality of equipment, the apparatus involving a processor, configured to detect, from time series sensor data received from the plurality of equipment, starting points across one or more operating cycles including a loading mode of operation, an unloading mode of operation, a stable mode of operation, and an off mode; extract a plurality of features from the time series sensor data for each of the loading mode of operation, the unloading mode of operation, the stable mode of operation, and the off mode in the one or more operating cycles based on the detected starting points; and generate one or more profiling models from the extracted features for the plurality of equipment.

[0012] Aspects of the present disclosure include a method for management of an installation of new equipment in a system involving a plurality of equipment, the method involving detecting, from time series sensor data received from the new equipment, new starting points for a loading mode of operation, an unloading mode of operation, a stable mode of operation, and an off mode; extracting a plurality of features from the time series sensor data for each of the loading mode of operation, the unloading mode of operation, the stable mode of operation, and the off mode based on the detected new starting points of the new equipment; running one or more profiling models on another plurality of features to generate a working probability; and for the working probability being below a threshold, indicating an improper installation of the new equipment.

[0013] Aspects of the present disclosure include a computer program for management of an installation of new equipment in a system involving a plurality of equipment, the computer program having instructions involving detecting, from time series sensor data received from the new equipment, new starting points for a loading mode of operation, an unloading mode of operation, a stable mode of operation, and an off mode; extracting a plurality of features from the time series sensor data for each of the loading mode of operation, the unloading mode of operation, the stable mode of operation, and the off mode based on the detected new starting points of the new equipment; running one or more profiling models on another plurality of features to generate a working probability; and for the working probability being below a threshold, indicating an improper installation of the new equipment. The instructions may be stored on a non-transitory computer readable medium and executed by one or more processors.

[0014] Aspects of the present disclosure include a system for management of an installation of new equipment in a system involving a plurality of equipment, the system involving means for detecting, from time series sensor data received from the new equipment, new starting points for a loading mode of operation, an unloading mode of operation, a stable mode of operation, and an off mode; means for extracting a plurality of features from the time series sensor data for each of the loading mode of operation, the unloading mode of operation, the stable mode of operation, and the off mode based on the detected new starting points of the new equipment; means for running one or more profiling models on another plurality of features to generate a working probability; and for the working probability being below a threshold, means for indicating an improper installation of the new equipment.

[0015] Aspects of the present disclosure involve an apparatus for management of an installation of new equipment in a system comprising a plurality of equipment, the apparatus involving a processor configured to detect, from time series sensor data received from the new equipment, new starting points for a loading mode of operation, an unloading mode of operation, a stable mode of operation, and an off mode; extract a plurality of features from the time series sensor data for each of the loading mode of operation, the unloading mode of operation, the stable mode of operation, and the off mode based on the detected new starting points of the new equipment; running one or more profiling models on another plurality of features to generate a working probability; and for the working probability being below a threshold, indicate an improper installation of the new equipment.

BRIEF DESCRIPTION OF DRAWINGS

[0016] FIG. 1 illustrates an example of a system architecture for machine learning based industrial equipment such as an air compressor working pattern profiling system that is built on top of the loT platform, a maintenance event database and an equipment transactional database, in accordance with an example implementation.

[0017] FIG. 2 illustrates an example of a working pattern profiling system as applied to an air compressor, in accordance with an example implementation.

[0018] FIG. 3 illustrates an example of the four different working modes of an air compressor, in accordance with an example implementation.

[0019] FIG. 4(a) illustrates a working mode segmentation method, in accordance with an example implementation.

[0020] FIG. 4(b) illustrates an example of the decision rules for the loading and off working modes, in accordance with an example implementation.

[0021] FIG. 4(c) illustrates an example of the decision rules for the stable and unloading working modes, in accordance with an example implementation.

[0022] FIG. 4(d) illustrates example sample features for each cycle, in accordance with an example implementation.

[0023] FIG. 5 illustrates an example flow diagram of profiling modeling for identifying an industry specific working pattern, in accordance with an example implementation.

[0024] FIG. 6 illustrates an example flow diagram for applying the profiling model to detect if the new equipment works normally, in accordance with an example implementation.

[0025] FIG. 7 illustrates an example configuration of an air compressor, in accordance with an example implementation.

[0026] FIG. 8 illustrates a system involving a plurality of air compressors networked to a management apparatus, in accordance with an example implementation.

[0027] FIG. 9 illustrates an example computing environment with an example computer device suitable for use in some example implementations. DETAILED DESCRIPTION

[0028] The following detailed description provides details of the figures and example implementations of the present application. Reference numerals and descriptions of redundant elements between figures are omitted for clarity. Terms used throughout the description are provided as examples and are not intended to be limiting. For example, the use of the term “automatic” may involve fully automatic or semi-automatic implementations involving user or administrator control over certain aspects of the implementation, depending on the desired implementation of one of ordinary skill in the art practicing implementations of the present application. Selection can be conducted by a user through a user interface or other input means, or can be implemented through a desired algorithm. Example implementations as described herein can be utilized either singularly or in combination and the functionality of the example implementations can be implemented through any means according to the desired implementations.

[0029] FIG. 1 illustrates an example of a system architecture for machine learning based industrial equipment such as an air compressor working pattern profiling system 100 that is built on top of the loT platform 120, a maintenance event database 124 and an equipment transactional database used to retrieve device attributes 125, in accordance with an example implementation. The industrial equipment loT platform includes sensors installed on each equipment 123 and a database to store sensor measurement 121 and operating control parameters 122, such as industry sector, geographical location, and so on.

[0030] Industrial equipment working pattern profiling server 110 can include data gathering and pre-processing 113, equipment working pattern profiling modeling 112, and profiling application modules 111. Data gathering and pre-processing module 113 first collects relevant sensor data as well as data from other sources such as equipment/device attributes 125 from a transactional database, and maintenance event data 124 from maintenance database. Then the collected data will be cleaned and consolidated to pass to the working pattern profiling module 112. Equipment working pattern profiling module 112 first analyzes operating sensor data and then apply machine learning algorithms to generate the corresponding pattern for a specific application. Profiling application module 111 includes, but is not limited to, applications such as industry insights configured to generate industry or process specific working patterns, anomaly detection configured to detect abnormal operating behavior from each air compressor through using normal operating pattern learned from data, energy efficiency recommendation configured to recommend the best operating parameters to improve energy efficiency, and compressor design recommendation. Client devices 130 can be utilized to review analytics from the loT platform 120 and can be utilized to select an industry of interest to conduct further analytics as described herein.

[0031] FIG. 2 illustrates an example of a working pattern profiling system 200 as applied to an air compressor, in accordance with an example implementation. Specifically, FIG. 2 illustrates a compressor working pattern profiling system 200 and a variety of methods contained therein.

[0032] The working pattern profiling system 200 intakes operational, maintenance, and transaction data 210, which can involve sensor measurements 211, operating control parameters 212, maintenance events 213, and device attributes 214 for one or more of the equipment managed by the working pattern profiling system. Such data is then pre-processed by data pre-process 220 to generate time series data for the modeling 230. Data pre-process 220 can include loT data formatting 221, sensor selection 222, functions to join loT data with non-IoT data 223 to link all the data in a time series manner, and device attribute selection 224. Sensor measurements 211 can involve raw data from the sensors of the equipment, which can be processed into sensor times series data through loT data formatting 221. Operating control parameters 212 indicate the parameter settings for the underlying equipment, from which the specific process or industry can be identified. For example, for an air compressor being implemented in the oil and gas industry, the operating control parameters 212 may involve various settings for fan control, discharge pressure, and so on to facilitate the implementation. Maintenance events 213 can be indicative of maintenance that occurred on the equipment, and can include maintenance logs. Device attributes 214 can involve attributes of the equipment, such as the equipment type, the model, and so on.

[0033] Working pattern profiling modeling module 230 includes components such as working mode segmentation 231, working mode based feature extraction 232, and profiling modeling 233 to generate one or more profiling models for the equipment managed by the system.

[0034] The profiling applications module 240 makes use of the models built in modeling module 230 to generate valuable insights or recommendations, including but not limited to industry insights 241 for a particular industry selected by client device 130, detecting abnormal behavior from a compressor 242, improving energy efficiency 243, assistance in design improvement 244, and other applications 245 in accordance with the desired implementation. In an example implementation for industry insights 241 and design improvement 244, the stored trained industry /process specific model can output the features determined to be important to guide the manufacturer to design specifics to satisfy the usage pattern determined from the industry insights 241.

[0035] In an example of a guideline for a new compressor with a similar process or in the same industry, because the new compressor does not have sufficient historical data yet, the built working pattern signature will help to detect if the new compressor is working normally. Further details of the application process are illustrated in FIG. 6. In addition, the profiling modeling procedure is illustrated in FIG. 5.

[0036] For working mode segmentation 231, based on the discharge air pressure of a working compressor, there can be various different working modes, such as loading mode in which the discharge pressure increases to reach the target level, stable mode in which the discharge pressure is relatively stable at the target level, unloading mode in which discharge pressure decreases to zero, and off mode in which discharge pressure is at zero or the compressor not working. The transition of modes from loading, to being stable, to unloading, then to off is referred to herein as a cycle or operating cycle.

[0037] From both energy efficiency and a compressor design point of view, it can be important to understand the compressor behavior/sensor values during each working mode. For example, it is important to understand how fast the compressor can load from zero discharge pressure to target pressures level.

[0038] However, it can be difficult to accurately separate these working modes due to the following reasons: i) during the loading/unloading period, the discharge pressure is not always monotonically increasing/decreasing; ii) during the stable period, the discharge pressure always goes up and down.

[0039] FIG. 3 illustrates an example of the four different working modes of an air compressor, in accordance with an example implementation. As illustrated in FIG. 3, the discharge pressure can have patterns indicative of the loading mode of operation, stable mode of operation, unloading mode of operation, and off mode. Moreover, in many situations in order to reduce cost, the compressor sensor values are collected at a crude resolution (e.g., every 30 minutes). When the working modes are segmented, if even one data point is missegmented, then there can be 30 minutes of discrepancy from the actual operating mode, which may affect the loading phase performance evaluation.

[0040] Working mode segmentation falls into the general time series segmentation problem. There are a variety of ways of formulating the segmentation problem in terms of defining the key parameters (number of segments, segmentation starting point, length of segments, error function, user-specified threshold, and so on). However, the universal algorithm (i.e., which in all cases results in optimal solution) does not exist. Further, related art implementations focus on the minimizing the approximation error that is defined by the Euclidian distance between the approximated Piecewise Linear Approximation and the original time series.

[0041] In contrast, the example implementations are directed to not reducing the approximation error, but to accurately identify the starting point of each working mode. Based on the specific working cycles of air compressors, the working mode segmentation method is developed as illustrated in FIG. 4(a).

[0042] FIG. 4(a) illustrates a working mode segmentation method, in accordance with an example implementation. Note that at the end of the flow at 404, not all data points during the “stable” mode segment will have a zero derivative. The flow at 405 and 406 intend to further reduce the small variations among data points within the “stable” mode segment. The decision rule for identifying the starting point of each segment is also illustrated in FIGS. 4(b) and 4(c).

[0043] At 400, the flow receives a sensor times series input for processing the working modes. In the example implementation involving air compressors, the sensor time series input is discharge pressure processed into time series, however, other data may also be used to determine working modes depending on the underlying equipment and the desired implementation, and the present disclosure is not limited thereto.

[0044] At 401 , the flow detects the “off’ segments first, which tend to have sensor readings of zero for applicable sensors. In the example of discharge pressure, the off segments are identified through identifying any continuous segment of zero discharge pressure.

[0045] At 402, the segments of the times series sensor data determined to be directed to the off mode is removed to form a series of disjoint “non-off’ segments. The start of each “non- off” segment is labeled as the start of “Loading” mode of operation and the end of each “non- off ’ segment is labeled as the start of “Off’ mode 402. The decision rules for the loading and off modes of operation are illustrated in FIG. 4(b).

[0046] At 403, for each “non-off” segment, the flow utilizes a filter to conduct smoothing of the data points. In an example implementation, a filter such as the Savitzky-Golay filter can be used to smooth the data points, however, any other localized filter can be utilized in accordance with the desired implementation, and the present disclosure is not limited thereto.

[0047] At 404, the flow determines the derivative of the smoothed time series. At 405, the flow utilizes a filter to conduct smoothing of the derivative of the smoothed time series. In an example implementation, a filter such as the Savitzky-Golay filter can be used to smooth the derivative, however, any other localized filter can be utilized in accordance with the desired implementation, and the present disclosure is not limited thereto.

[0048] At 406, from the smoothed derivative in the flow of 405, the flow replaces the derivative data within +/- one standard deviation of the mean with 0 to further reduce noise. Such derivatives will be used for further processing in the flow as described below.

[0049] At 407, the flow labels the first data point having a derivative of zero while the derivative at the previous point is positive as the starting point of the “Stable” mode of operation; and also labels the first data point having a derivative of zero while the derivative at the previous point is negative as the starting point of the “Unloading” mode of operation. The decision rules for the unloading and the stable modes of operation are illustrated in FIG. 4(c).

[0050] At 408, the flow output the starting points of all the modes for each segment: Loading, Stable, Unloading, Off. Through the example implementations described herein, sliding windows can thereby be avoided for determining modes of operations, as sliding windows may mix modes of operations together.

[0051] FIG. 4(b) illustrates an example of the decision rules for the loading and off working modes of operation, in accordance with an example implementation. In the example of FIG. 4(b), the decision rules for the loading and off working modes of operation are directed to an air compressor through the flow of FIG. 4(a), however, such rules can be adjusted for any other equipment in accordance with the desired implementation with the appropriate sensor data. [0052] FIG. 4(c) illustrates an example of the decision rules for the stable and unloading working modes, in accordance with an example implementation. In the example of FIG. 4(c), the decision rules for the stable and unloading working modes of operation are directed to an air compressor through the flow of FIG. 4(a), however, such rules can be adjusted for any other equipment in accordance with the desired implementation with the appropriate sensor data.

[0053] For working mode based feature extraction 232, after all of the working modes are detected for the whole operating period, features can be extracted for each working mode segment within a cycle from different sensor values. Compared with the traditional sliding window or landmark window feature extraction, working mode based feature extraction offers a more meaningful way of feature extraction.

[0054] FIG. 4(d) illustrates example sample features for each cycle, in accordance with an example implementation. The example sample features that can be extracted for each cycle as listed in FIG. 4(d) are directed to air compressors, however, such features can be modified for any other equipment in accordance with the desired implementation.

[0055] For profiling modeling 233, instead of using ad-hoc profiling by selecting a few plausible sensor values based on domain knowledge or intuition, example implementations described herein involve data driven machine learning algorithms to profile working patterns. That is, depending on the target application, application-specific machine learning models are constructed.

[0056] FIG. 5 illustrates an example flow diagram of profiling modeling for identifying an industry specific working pattern, in accordance with an example implementation. Although the example of FIG. 5 is directed to compressors, other equipment may be utilized as well and the present disclosure is not limited thereto.

[0057] The flow takes in the feature table generated from all extracted equipment of the type of interest from working mode base feature extraction module as input at 500. At 501, the flow receives a selection of a target industry of interest I. In an example involving air compressors, such air compressors can be deployed for use in various industries (e.g., oil and gas, HVAC systems, and so on), and all managed by the same loT platform. The flow thereby obtains the air compressors that are assigned for use at the selected industry of interest. At 502, the flow retrieves the equipment (e.g., compressors) that are used in industry I. Suppose such equipment has worked for n cycles. Denote the feature matrix as:

[0059] At 503, the flow randomly select n cycles from equipment used in any other industry, denote the features for these randomly selected cycles as:

[0060]

[0061] At 504, the flow combines the feature matrices to generate a combined matrix as

[0063] The flow creates a label vector of length 2n with the first n elements as “1” and the rest of the n elements as “0”. Denote the label vector as Y=

[1, i, . „ 1,0, 0. 0] ^T

[0064] At 505, the flow trains and cross-validate a binary classifier (e.g., support vector machine (SVM), RandomForest, artificial neural network (ANN), and so on) using the combined feature matrix from 504, and label vector Y.

[0065] At 506, the flow stores the classifier model with the best performance for industry I working pattern.

[0066] At 507, a determination is made as to whether there are other industries that are to be analyzed. If so (Yes), then the flow proceeds to 501, otherwise (No), the flow ends.

[0067] Some model building examples are illustrated below.

[0068] In an example for determining an industry /process specific working pattern signature of a certain compressor type, there is a need to understand the working pattern of a certain compressor type for a specific industry in order to help the compressor design. Example implementations first extract a set of compressors of the type of interest that are used in multiple industries/processes. By applying the working mode segmentation module 231 and the working mode based feature extraction module 232 to each of the compressor historical sensor values, the example implementations generate a set of training data containing the extracted features.

[0069] In an example for determining the compressor normal operating signature, example implementations can also build a normal operating pattern signature for each individual compressor. The process is similar as in the industry/process specific working pattern signature determination, except that the positive training data come from the normal operating data of the specific compressor. The negative training data can come from the abnormal operating period of the compressor or operating data from other compressors with industries/types.

[0070] In an example of energy efficiency modeling, the model can be built to analyze how working patterns affect energy consumptions in accordance with an example implementation. The energy consumption may be measured by electric power in watts. Then, the example implementations can build regression models using the energy consumption as a target value and the features extracted as input variables. The model then outputs the relationship between the energy consumption and working patterns.

[0071] FIG. 6 illustrates an example flow diagram for applying the profiling model to detect if the new equipment works normally, in accordance with an example implementation. In the example of FIG. 6, the example involves an installation of a new compressor that is connected to the loT platform, however, other equipment can also be connected and the present disclosure is not limited thereto.

[0072] At 600, the flow takes in sensor data from a new compressor used in an industry or specific process as an input. At 601, the flow executes the Working Mode Segmentation Module to segment each working cycle into the Loading mode of operation, the Stable mode of operation, the Unloading mode of operation, and the Off mode based on the sensor data. In the example of air compressor, the modes of operation and determined from the discharge pressure time series, however, such example implementations can be adjusted based on the underlying equipment in accordance with the desired implementation.

[0073] At 602, the flow runs the working mode based feature extract module to extract features from all types of sensor values. At 603, the flow retrieves the model signature for the industry or specific process built in the profiling modeling module and supplies the features generated for the new compressor to the retrieved model. The model then outputs a similarity probability. At 604, the if the probability is high (e.g., greater than 0.5), then the model indicates that the new equipment is working similarly as standard working pattern in that industry or with the process. Otherwise, the flow may require some attention to check if the new equipment is working properly.

[0074] With respect to the anomaly detection application of compressor normal operating signature 242, example implementations can feed in the new operating data to mode segmentation module and the feature extraction module, then to the built normal operating signature to detect any abnormal behavior of the compressor.

[0075] With respect to the energy efficiency recommendation 243, the model built using energy consumption as the target KPI highlights the most contributing factors to energy consumption and provides recommendations on what are the best control parameters to save energy.

[0076] Through example implementations described herein, there can be a system and plurality of methods to analyze vast amount of industrial equipment loT data collected by manufacturers to build machine learning based working pattern profiling and applications to improve equipment design, equipment heath, energy efficiency, and so on.

[0077] FIG. 7 illustrates an example configuration of an air compressor, in accordance with an example implementation. Example components of an air compressor can include, but are not limited to an engine motor 19, compressor cooling system 15, receiver tank 16, and instrument panel/on board computer 1, which can be interconnected by an electrical system. Depending on the desired implementation, the compressor can also be supplied with sound deadening insulation to lower noise emissions to meet specified requirements.

[0078] The compressor cooling system 15 can involve components such as a radiator, high capacity fan, and thermostats. The high capacity fan draws air through the radiator, keeping the engine motor 19 at the desired operating temperature. The same fan also cools the fluid in the compressor cooling system 15. While passing through the radiator, the fan air also passes through the compressor fluid cooler. As air passes through the cooler, the heat of compression is removed from the fluid. The same fan also cools the engine intake air supply. While passing through the radiator, the fan air passes through an air to air aftercooler. As air passes through the air to air aftercooler, heat is removed from the engine motor 19. [0079] In example implementations of an air compressor, fluid is injected into the compressor unit and mixes directly with the air as the rotors turn, compressing the air. The fluid flow has several functions. As coolant, the fluid flow controls the rise of air temperature normally associated with the heat of compression. The fluid flow further seals the clearance paths between the rotors and the stator and also between the rotors themselves. The fluid flow also acts as a lubricating film between the rotors allowing one rotor to directly drive the other.

[0080] After the air/fluid mixture is discharged from the compressor unit, the fluid is separated from the air. At this time, the air flows through an aftercooler and water separator (if equipped) then to a service line while the fluid is being cooled in preparation for reinjection.

[0081] In example implementations of an air compressor, the air compressor discharges compressed air/fluid mixture into the receiver tank 16, from which the discharge pressure can be measured through appropriately placed sensors. The receiver tank 16 has several functions including acting as a primary fluid separator, serving as the compressor fluid storage sump and housing the final fluid separator.

[0082] The compressed air/fluid mixture enters the receiver tank 16 and is directed against the tank side wall. By change of direction and reduction of velocity, large droplets of fluid separate and fall to the bottom of the receiver tank. The fractional percentage of fluid remaining in the compressed air collects on the surface of the final separator element as the compressed air flows through the separator. As more and more fluid collects on the element’s surface, the fluid descends to the bottom of the separator. A return line (or scavenge tube) leads from the bottom of the separator element to the inlet region of the air compressor. Fluid collecting on the bottom of the separator element is returned to the compressor by the pressure difference between the area surrounding the separator element and the compressor inlet.

[0083] FIG. 8 illustrates a system involving a plurality of air compressors networked to a management apparatus, in accordance with an example implementation. One or more air compressors 801 are communicatively coupled to a network 800 (e.g., local area network (LAN), wide area network (WAN)) through the corresponding on-board computer of the air compressor 801, which is connected to a management apparatus 802. The management apparatus 802 manages a database 803, which contains historical data collected from the air compressors from each of the air compressors 801 and also facilitates remote control to each of the air compressors 801. In alternate example implementations, the data from the air compressors 801 can be stored to a central repository or central database such as proprietary databases that intake data from air compressors, or systems such as enterprise resource planning systems, and the management apparatus 802 can access or retrieve the data from the central repository or central database. In this manner, the management apparatus 802 can facilitate functionality of an analytics platform for the air compressors 801 or other equipment, and can also be configured to control or manage new installation of equipment.

[0084] FIG. 9 illustrates an example computing environment with an example computer device suitable for use in some example implementations, such as a management apparatus 802 as illustrated in FIG. 8, or as an on-board computer 1 as illustrated in FIG. 7. Computer device 905 in computing environment 900 can include one or more processing units, cores, or processors 910, memory 915 (e.g., RAM, ROM, and/or the like), internal storage 920 (e.g., magnetic, optical, solid state storage, and/or organic), and/or I/O interface 925, any of which can be coupled on a communication mechanism or bus 930 for communicating information or embedded in the computer device 905. I/O interface 925 is also configured to receive images from cameras or provide images to projectors or displays, depending on the desired implementation.

[0085] Computer device 905 can be communicatively coupled to input/user interface 935 and output device/interface 940. Either one or both of input/user interface 935 and output device/interface 940 can be a wired or wireless interface and can be detachable. Input/user interface 935 may include any device, component, sensor, or interface, physical or virtual, that can be used to provide input (e.g., buttons, touch-screen interface, keyboard, a pointing/ cursor control, microphone, camera, braille, motion sensor, optical reader, and/or the like). Output device/interface 940 may include a display, television, monitor, printer, speaker, braille, or the like. In some example implementations, input/user interface 935 and output device/interface 940 can be embedded with or physically coupled to the computer device 905. In other example implementations, other computer devices may function as or provide the functions of input/user interface 935 and output device/interface 940 for a computer device 905.

[0086] Examples of computer device 905 may include, but are not limited to, highly mobile devices (e.g., smartphones, devices in vehicles and other machines, devices carried by humans and animals, and the like), mobile devices (e.g., tablets, notebooks, laptops, personal computers, portable televisions, radios, and the like), and devices not designed for mobility (e.g., desktop computers, other computers, information kiosks, televisions with one or more processors embedded therein and/or coupled thereto, radios, and the like).

[0087] Computer device 905 can be communicatively coupled (e.g., via I/O interface 925) to external storage 945 and network 950 for communicating with any number of networked components, devices, and systems, including one or more computer devices of the same or different configuration. Computer device 905 or any connected computer device can be functioning as, providing services of, or referred to as a server, client, thin server, general machine, special-purpose machine, or another label.

[0088] I/O interface 925 can include, but is not limited to, wired and/or wireless interfaces using any communication or I/O protocols or standards (e.g., Ethernet, 802.1 lx, Universal System Bus, WiMax, modem, a cellular network protocol, and the like) for communicating information to and/or from at least all the connected components, devices, and network in computing environment 900. Network 950 can be any network or combination of networks (e.g., the Internet, local area network, wide area network, a telephonic network, a cellular network, satellite network, and the like).

[0089] Computer device 905 can use and/or communicate using computer-usable or computer-readable media, including transitory media and non-transitory media. Transitory media include transmission media (e.g., metal cables, fiber optics), signals, carrier waves, and the like. Non-transitory media include magnetic media (e.g., disks and tapes), optical media (e.g., CD ROM, digital video disks, Blu-ray disks), solid state media (e.g., RAM, ROM, flash memory, solid-state storage), and other non-volatile storage or memory.

[0090] Computer device 905 can be used to implement techniques, methods, applications, processes, or computer-executable instructions in some example computing environments. Computer-executable instructions can be retrieved from transitory media, and stored on and retrieved from non-transitory media. The executable instructions can originate from one or more of any programming, scripting, and machine languages (e.g., C, C++, C#, Java, Visual Basic, Python, Perl, JavaScript, and others).

[0091] Processor(s) 910 can execute under any operating system (OS) (not shown), in a native or virtual environment. One or more applications can be deployed that include logic unit 960, application programming interface (API) unit 965, input unit 970, output unit 975, and inter-unit communication mechanism 995 for the different units to communicate with each other, with the OS, and with other applications (not shown). The described units and elements can be varied in design, function, configuration, or implementation and are not limited to the descriptions provided. Processor(s) 910 can be in the form of hardware processors such as central processing units (CPUs) or in a combination of hardware and software units.

[0092] In some example implementations, when information or an execution instruction is received by API unit 965, it may be communicated to one or more other units (e.g., logic unit 960, input unit 970, output unit 975). In some instances, logic unit 960 may be configured to control the information flow among the units and direct the services provided by API unit 965, input unit 970, output unit 975, in some example implementations described above. For example, the flow of one or more processes or implementations may be controlled by logic unit 960 alone or in conjunction with API unit 965. The input unit 970 may be configured to obtain input for the calculations described in the example implementations, and the output unit 975 may be configured to provide output based on the calculations described in example implementations.

[0093] Processor(s) 910 can be configured to control the air compressors remotely through communication of instructions to a corresponding on-board computer of an air compressor. Such instructions can include, but are not limited to, power down, power up, engaging a maintenance mode, and so on in accordance with a desired implementation.

[0094] In an example of a management apparatus to facilitate an analytics platform, processor(s) 910 can be configured to detect, from time series sensor data received from the plurality of equipment, starting points across one or more operating cycles involving a loading mode of operation, an unloading mode of operation, a stable mode of operation, and an off mode; extract a plurality of features from the time series sensor data for each of the loading mode of operation, the unloading mode of operation, the stable mode of operation, and the off mode in the one or more operating cycles based on the detected starting points; and generate one or more profiling models from the extracted features for the plurality of equipment.

[0095] Processor(s) 910 can be configured to detect, from the time series sensor data received from the plurality of equipment, the starting points for the loading mode of operation, the unloading mode of operation, a stable mode of operation, and the off mode by detecting first segments in the time series sensor data having substantially zero continuous sensor readings; removing the first segments from the time series sensor data to form second segments; identifying a start of each of the second segments as the starting points for the loading mode of operation; generating and smoothing derivatives of the second segments; identifying first points in the smoothed derivatives of the second segments having a zero derivative with a positive previous point as the starting points for the stable mode; and identifying second points in the smoothed derivatives of the second segments having a zero derivative with a negative previous point as the starting points for the unloading mode as illustrated in FIG. 4.

[0096] For receipt of a target industry of interest, processor(s) 910 can be configured to retrieve ones of the plurality of equipment operating in the target industry of interest; generate combined features from the extracted features associated with the ones of the plurality of equipment and a labeling vector; generate a plurality of machine learning classifier models from the combined features and the labeling vector; and store one of the plurality of machine learning classifier models having a highest performance for the target industry of interest as illustrated in FIG. 5.

[0097] For an installation of new equipment in the plurality of equipment, the processor is configured to detect, from time series sensor data received from the new equipment, new starting points for the loading mode of operation, the unloading mode of operation, the stable mode of operation, and the off mode; extract another plurality of features from the time series sensor data for each of the loading mode of operation, the unloading mode of operation, the stable mode of operation, and the off mode based on the detected new starting points of the new equipment; run the one or more profiling models on the another plurality of features to generate a working probability; and for the working probability being below a threshold, indicate an improper installation of the new equipment as illustrated in FIG. 6.

[0098] As illustrated in FIG. 7, the plurality of equipment can involve air compressors, although other equipment can be utilized in accordance with the desired implementation.

[0099] In an example of a management apparatus configured to facilitate an installation of new equipment in a system involving a plurality of equipment, processor(s) 910 can be configured to detec, from time series sensor data received from the new equipment, new starting points for a loading mode of operation, an unloading mode of operation, a stable mode of operation, and an off mode; extract a plurality of features from the time series sensor data for each of the loading mode of operation, the unloading mode of operation, the stable mode of operation, and the off mode based on the detected new starting points of the new equipment; run one or more profiling models on another plurality of features to generate a working probability; and for the working probability being below a threshold, indicating an improper installation of the new equipment.

[0100] Example implementations are not limited to industrial equipment such as compressor design, equipment user, equipment operator, but can be extended to other equipment depending on the desired implementation. The example implementations can be used as a standalone solution or be integrated with existing loT systems.

[0101] Some portions of the detailed description are presented in terms of algorithms and symbolic representations of operations within a computer. These algorithmic descriptions and symbolic representations are the means used by those skilled in the data processing arts to convey the essence of their innovations to others skilled in the art. An algorithm is a series of defined steps leading to a desired end state or result. In example implementations, the steps carried out require physical manipulations of tangible quantities for achieving a tangible result.

[0102] Unless specifically stated otherwise, as apparent from the discussion, it is appreciated that throughout the description, discussions utilizing terms such as “processing,” “computing,” “calculating,” “determining,” “displaying,” or the like, can include the actions and processes of a computer system or other information processing device that manipulates and transforms data represented as physical (electronic) quantities within the computer system’s registers and memories into other data similarly represented as physical quantities within the computer system’s memories or registers or other information storage, transmission or display devices.

[0103] Example implementations may also relate to an apparatus for performing the operations herein. This apparatus may be specially constructed for the required purposes, or it may include one or more general-purpose computers selectively activated or reconfigured by one or more computer programs. Such computer programs may be stored in a computer readable medium, such as a computer-readable storage medium or a computer-readable signal medium. A computer-readable storage medium may involve tangible mediums such as, but not limited to optical disks, magnetic disks, read-only memories, random access memories, solid state devices and drives, or any other types of tangible or non-transitory media suitable for storing electronic information. A computer readable signal medium may include mediums such as carrier waves. The algorithms and displays presented herein are not inherently related to any particular computer or other apparatus. Computer programs can involve pure software implementations that involve instructions that perform the operations of the desired implementation.

[0104] Various general-purpose systems may be used with programs and modules in accordance with the examples herein, or it may prove convenient to construct a more specialized apparatus to perform desired method steps. In addition, the example implementations are not described with reference to any particular programming language. It will be appreciated that a variety of programming languages may be used to implement the teachings of the example implementations as described herein. The instructions of the programming language(s) may be executed by one or more processing devices, e.g., central processing units (CPUs), processors, or controllers.

[0105] As is known in the art, the operations described above can be performed by hardware, software, or some combination of software and hardware. Various aspects of the example implementations may be implemented using circuits and logic devices (hardware), while other aspects may be implemented using instructions stored on a machine-readable medium (software), which if executed by a processor, would cause the processor to perform a method to carry out implementations of the present application. Further, some example implementations of the present application may be performed solely in hardware, whereas other example implementations may be performed solely in software. Moreover, the various functions described can be performed in a single unit, or can be spread across a number of components in any number of ways. When performed by software, the methods may be executed by a processor, such as a general purpose computer, based on instructions stored on a computer-readable medium. If desired, the instructions can be stored on the medium in a compressed and/or encrypted format.

[0106] Moreover, other implementations of the present application will be apparent to those skilled in the art from consideration of the specification and practice of the teachings of the present application. Various aspects and/or components of the described example implementations may be used singly or in any combination. It is intended that the specification and example implementations be considered as examples only, with the true scope and spirit of the present application being indicated by the following claims.