Building real-time FDA compliant biomedical wearable products: challenges and commercial solutions

,

Recent research suggests that ECG wearables devices (such as smart watches) are now medically suitable for providing predictive insights into serious heart conditions such as atrial fibrillation (A-Fib). These advancements have been facilitated by the availability of low-cost microcontrollers offering algorithmic functionality, allowing developers to implement wearables with excellent battery life and edge-based real-time data analysis.

Although the international research community has produced many innovative high-performance ECG and PPG biomedical algorithms, these are unfortunately limited to offline clinical analysis in Matlab or Python. As such, very little emphasis has been placed on building commercial real-time wearables algorithms on microcontrollers, leading manufacturers to conduct the research themselves and to design suitable candidates. 

This is further complicated by the requirement of manufacturers on how they will implement a developed algorithm in real-time on a low-cost microcontroller and still achieve decent battery life.

Arm Cortex-M microcontrollers

Over 90% of the microcontrollers used in the smart product market are powered by so-called Arm Cortex-M processors that offer a combination of high algorithmic performance, low-power and security. The Arm Cortex-M4 is a very popular choice with hundreds of silicon vendors (including ST, TI, NXP, ADI, Nordic, Microchip, Renesas), as it offers DSP (digital signal processing) functionality traditionally found in more expensive devices and is low-power.

The Cortex-M4F device offers floating point support, helping with RAD (rapid application development) as designs can be easily ported from Matlab/Python to C without the need of performing a detailed quantisation arithmetic analysis. As such, a design cycle can be cut from months to weeks, offering organisations a significant cost saving.

Arm and its rich ecosystem of partners provide developers with easy-to-use tooling and tried and tested software libraries, such as the CMSIS-DSP and CMSIS-NN frameworks and ASN’s DSP filtering library for algorithm development and machine learning.

FDA compliance

The AHA (American Heart Association) provides developers with guidelines for developing FDA-compliant ECG monitoring products. These are broken down into the following three categories: 

  1. Diagnostic: 0.05Hz -150Hz
  2. Ambulatory (wearables): 0.67Hz – 40Hz
  3. ST segment: 0.05Hz

The ECG measurements must be FDA compliant with IEC 60601-2 2-47 standards for ambulatory ECG, but what are the criteria and challenges?

Challenges with ECG/PPG measurements

Modelling the QRS complex found in ECG data is extremely difficult, as to date there is no concrete model available.  This is further complicated by the variety of ECG data depending on the position of the lead on the patient’s body and illnesses. The following list summaries the typical challenges faced by algorithm developers:

  1. Accurate baseline wander (BLW) removal remains one of the most challenging topics in ECG analysis.
  2. The BLW must be removed for accurate clinical analysis.
  3. BLW manifests itself as low-frequency ‘wander’ (typically <0.5Hz) from EMG and torso movement.
  4. QRS width widening and amplitude distortion due to filtering invalidates clinical analysis.
  5. Reducing EMG and measurement noise without altering the temporal biomedical relationships of the ECG signal.
  6. 50/60Hz powerline interference can swamp the ECG signal – this is primarily attributed to pickup by the long high impedance measurement cables. This is typically problematic for extended bandwidth wearable applications that go beyond 40Hz.
  7. Glitches, sudden movement and poor sensor contact with the skin: This is related to BLW, but usually manifests itself as abrupt glitches in the ECG measurement data. The correction algorithm must discriminate between these undesirable events and normal behaviour.
  8. IEC 60601-2 2-47 frequency response specifications:
    • Bandwidth: 0.67Hz – 40Hz.
    • Passband ripple: < ±0.5dB
    • Maximum ±10% amplitude error: most biomedical SoCs make use of a Sigma-Delta ADC, leading to amplitude droop.

Shortcomings with ECG/PPG algorithms

A mentioned in the previous section, much research has been conducted over the years with mixed results. The main shortcomings of these methods are summarised below:

  1. Computationally heavy: most algorithms have been designed for research in Matlab and not for real-time, e.g. wavelets have excellent performance but have high computational cost, leading to poor battery life and the need for an expensive processor.
  2. Large latency and warping: digital filtering chain introduces large latency, computational cost and can warp the characteristics of the biomedical features.
  3. Overlapping frequencies: there are many examples of unwanted noise overlapping the delicate ECG data, hence the popularity of time-frequency analysis, such as wavelets.
  4. Mixed results regarding BLW removal: spline removal is excellent, but it has high computational cost and has the added difficulty of finding good correction points between the QRS complexes. Linear phase FIR filtering is a good compromise but has very high computational cost (typically >1000 filter coefficients) due to the high sampling rate to cut-off frequency ratio. Non-linear phase IIR filter has low computational cost, but warps the ECG features, and is therefore unsuitable for clinical analysis.
  5. AI based kernel filters: ‘black box filter’ based on massive training data. Moderate implementation cost with performance dependent on the variety of training data, leading to unpredictable results in some cases.
  6. PPG analysis: has the added difficulty of eliminating motion from the measurement data, such as when walking or running. Although a range of tentative algorithms has been proposed by various researchers using accelerometer measurement data to correct the PPG data, very few commercial solutions are currently based on this technology.

It would seem that ECG and PPG analysis has some major obstacles to overcome, especially when considering how to deploy the algorithms on low-power microcontrollers.

The future: ASN’s real-time RCF algorithm and Advanced Analytics

Together with cardiologists from Medisch Spectrum Twente, ASN’s advanced analytics team developed the RCF (retrospective collaborative filtering) algorithm that uses time-frequency analysis to enhance the ECG data in real-time.

The essence of RCF algorithm centres around a highly optimised set of polynomial cleaning filters with different frequency characteristics that are applied to different segments of the QRS complex for enhancement. This has some synergy with wavelets, but it does not suffer from the computational burden associated with wavelet analysis.

The polynomial filters are peak preserving, meaning that they preserve the delicate biomedical peaks while smoothing out the unwanted noise/ripple. The polynomial fitting operation also overcomes the challenge of overlapping frequency content, as data within a specified region is smoothed out by the relevant filter.

RCF is further strengthened by the BLW killer IP block that implements a highly computationally efficient linear phase 0.67Hz highpass filter. The net effect is an FDA-compliant signal chain suitable for clinical analysis. The complete signal chain is extremely computationally efficient, and as such is suitable for Arm’s popular M3 and M4 Cortex-M processor families.

Real-time ECG feature extraction

The ECG waveform can be split up into segments, where each wave or segment represents a certain event in the cardiac cycle, as shown below:

As seen, the biomedical features are designated P, Q, R, S, T that define points in time within the cardiac cycle. The RCF algorithm is further strengthened with our state-of-the-art AAE (Advanced Analytics Engine) that automatically cleans and find these features for clinical analysis.

AAE supported analytics

  1. P-wave duration
  2. PR interval
  3. QRS duration
  4. QT duration (Bazett algorithm used for QTc)
  5. HR (RR interval)
  6. HRV (rMSSD algorithm used)

Armed with the real-time features, an ML model can be trained and provide valuable insights into patient health running on an edge processor inside a wearable device.

A-Fib

Atrial fibrillation (A-Fib) is the most frequent cardiac arrhythmia, affecting millions of people worldwide. An arrhythmia is when the heart beats too slowly, too fast, or in an irregular way. Signs of A-Fib are an irregular beating pattern and no p-waves. Our AAE provides developers with all of the relevant features needed to build an ML model for robust A-Fib detection.

Let us help you build your product

By combining advanced low-power processor technology, advanced mathematical algorithmic concepts and medical knowledge, our solution provides developers with an easy way of building wearable products for medical use. The high accuracy of our Advanced Analytics Engine (AAE) has been verified by cardiologists, and can be used with an additional ML model or standalone to provide people with valuable insights into potentially fatal health conditions, such as A-Fib without the need for an expensive medical examination at a hospital.

ASN’s ECG algorithmic solutions are ideal for building next generation ECG and PPG wearable products on Arm Cortex-M microcontrollers (e.g. STM32F4, MAX32660) and bio-sensor SoCs (MAX86150). These algorithms can be easily used with industry standard biomedical AFEs, such as: MAX30003, AFE4500 and AFE4950.

Please contact us for more information and to arrange an evaluation.

Author

  • Sanjeev is a RTEI (Real-Time Edge Intelligence) visionary and expert in signals and systems with a track record of successfully developing over 25 commercial products. He is a Distinguished Arm Ambassador and advises top international blue chip companies on their AIoT/RTEI solutions and strategies for I4.0, telemedicine, smart healthcare, smart grids and smart buildings.

    View all posts
/by

Deploying AI based CbM applications: challenges and algorithmic solutions

,

Unexpected equipment failures can be expensive and potentially catastrophic, resulting in unplanned production downtime, costly replacement of parts and safety and environmental concerns. With many factories and process control plants facing an ever-increasing shortage of experienced personnel, many are now looking for AI based systems to replace the ‘experienced old guy’ who knows everything about the machine and reduce their Total Cost of Ownership (TOC).

The challenge is however, how do you build and train an AI CbM system to replace an expert ?

What is CbM?

As part of the I4.0 revolution, Condition based monitoring (CbM) of machines has received a great amount of attention, as factories look to maximise their production efficiency and reduce their TOC, while at the same time retaining the invaluable skills of experienced foremen and production workers. As such, CbM is a process for monitoring equipment during operation to identify any deterioration, enabling maintenance to be planned and operational costs reduced.

CbM 5G edge computing

Many are factory owners are suspicious of cloud-based enterprise solutions offered by Microsoft, Amazon and Google as data leaves the site and any latency issues could affect production output. Recently 5G edge computing has received much attention, whereby all time-critical operations are undertaken at the edge (i.e. near to the asset in the factory) via smart sensors.

Arm’s rich set of Cortex processors offer a combination of high performance, ML/DSP computation support and low power. This is further strengthened by Arm’s new Helium Cortex-M55 and Cortex-M85 AI based processors that have been specially designed for edge-based AI applications – the latter offers an impressive 3DMIPS/MHz making it a good fit for ML and DSP algorithms. These processors and supporting libraries now allow developers to develop high-performance CbM smart sensors to perform their computationally intensive tasks at the edge and communicate the results via a 5G network to a smartphone or database. This provides higher reliability and scalability than expensive cloud-based solutions reliant on big data.

It would seem that big data has had its day!

Vibration sensor technology

Contactless MEMS (microelectromechanical systems) accelerometers sensors are an excellent alternative to the well-established, but bulky and expensive (25-500+ EUR) Piezo sensors for obtaining vibration information. MEMS sensors are relatively low cost (10-30 EUR) and can offer a response down to DC (zero Hertz), which is useful for the detection of imbalance at very low rotational speeds. MEMS accelerometers also have a self-test feature whereby the sensor can be verified to be 100% functional. They produce acceleration data that can be analysed by various vibration monitoring algorithms.

Spectral vibration monitoring via the FFT (Fast Fourier Transform) is regarded as an industry standard for machine vibration analysis. If a mechanical problem exists, the FFT spectra (multiple spectrums) will provide information to help determine the source and cause of the problem. Coupled with the right AI algorithms, the features from the FFT analysis can be used to identify the root cause of the failure, such as motor imbalance, misalignment, and looseness. These properties and challenges faced by the FFT will be discussed further later on in the article.

There are several steps to follow as guidelines to help achieve a successful vibration monitoring programme. The following is a general list of these steps:

  • Collect useful information: Look, listen and feel the machinery to check for resonance. Identify what measurements are needed (point and point type). Conduct additional testing if further data are required.
  • Analyse spectral data: Evaluate the overall values and specific frequencies corresponding to machinery anomalies. Compare overall values in different directions and current measurements with historical data.
  • Multi-parameter monitoring: Use additional techniques to conclude the fault type. (Analysis tools such as phase measurements, current analysis, acceleration enveloping, oil analysis and thermography can also be used.)
  • Perform Root Cause Analysis (RCA): In order to identify the real causes of the problem and to prevent it from occurring again.
  • Reporting and planning actions: Use a Computer Maintenance Management System (CMMS) to rectify the problem and take action to achieve a plan.

Getting acceleration, velocity and/or displacement estimates

As aforementioned, a popular device used to obtain acceleration data is a so-called ‘accelerometer’. These devices are semiconductor-based MEMS (microelectromechanical systems) and provide 3D (i.e. tri-axial) acceleration time domain data to a supporting microcontroller.

Before FFT analysis, the accelerometer data is usually passed through integration signal processing blocks, in order to convert the time domain acceleration data into velocity and displacement data. These blocks are comprised of a highpass filter and cumulative sum (integration). The highpass filter is essential for removing the effects of DC and noise, which would cause an offset in the output (i.e. the result of the integration). Depending on the severity of the noise/DC the output may even saturate, making it unusable for analysis. The design of a suitable highpass filter is an extremely challenging task and is the primary reason why many vibration analysis systems struggle to measure vibrations <10Hz (600 RPM).

Collect useful information

When conducting a vibration program, certain preliminary information is needed in order to conduct an analysis. The identification of components, running speed, operating environment and types of measurements should be determined initially to assess the overall system.

Identify components of the machine that could cause vibration

Before a spectrum can be analysed, the components that cause vibration within the machine must be identified. For example, you should be familiar with these key components:

  • If the machine is connected to a fan or pump, it is important to know the number of fan blades or impellers.
  • If bearings are present, know the bearing identification number or its designation.
  • If the machine contains, or is coupled, to a gearbox, know the number of teeth and shaft speeds.
  • If the machine is driven with belts, know the belt lengths.

The above information helps assess spectral components and helps identify the vibration source. Determining the running speed is the initial task. There are several methods to help identify this parameter.

Identifying the running speed

Knowing the machine’s running speed is critical when analysing an FFT spectrum. Running speed is related to most components within the machine and therefore, aids in assessing overall machine health. There are several ways to determine running speed:

  • Read the speed from instrumentation at the machine or from instrumentation in the control room monitoring the machine.
  • Look for peaks in the spectrum at 1,800 or 3,600 RPM (60Hz countries), 1,500 and 3,000 RPM (50Hz countries) if the machine is an induction electric motor, as electric motors usually run at these speeds. If the machine is variable speed, look for peaks in the spectrum that are close to the running speed of the machine during the time at which the data is captured.
  • An FFT’s running speed peak is typically the first significant peak in the spectrum when reading the spectrum from left to right. Search for this peak and check for peaks at two times, three times, four times, etc. (at the harmonic frequencies).

Challenges with the FFT algorithm

FFT spectra allow us to analyse vibration amplitudes at various component frequencies on the FFT spectrum. In this way, we can identify and track vibration occurring at specific frequencies. Since we know that particular machinery problems generate vibration at specific frequencies, we can use this information to diagnose the cause of excessive vibration.

Challenges with spectral analysis

  • The sampling rate of the accelerometer drifts with temperature: This results in a mismatch between the FFT analysis sampling frequency and the real situation. As such, the amplitude and frequency estimates of the vibration will be incorrect.
  • Frequency resolution: the frequency of the vibration peak may have a fractional value. If the resolution of the Fourier algorithm is not fine enough, it will ‘smear’ the result, resulting in a lower amplitude estimate.
  • Running speed: this is typically known apriori, but will have a degree of error associated with it and will change with temperature. For example, 3000 rpm ±1% is 50Hz ±0.5Hz at the fundamental running frequency. In order to track higher harmonics (i.e. multiples of the running speed) the FFT must have sufficient frequency resolution to accurately estimate the amplitude at the right frequency.

Traditional FFT based analysis uses a very high number of computational points in order to achieve a 1Hz resolution. Although this is OK, it still does not overcome the fractional frequency components and requires considerable computational effort.

Some designs use a phaselocked loop, that tracks the running frequency and sets the FFT analysis sampling frequency to a multiple (e.g. 20x) of the running speed. Although this is a very good workaround, it requires specialised hardware (such as an expensive ASIC) and is inflexible for changes in running speed.

ML feature extraction, DSP algorithms and models

In order to build an ML (machine learning) model for an AI CbM application, several challenges need to be overcome.

  • Definition of classes: In order to make a classification, ML classes must be defined. In the simplest sense, this can be Fault or Normal behaviour, but what about other cases?
  • ML Features: what data features will be used for the ML model? Running speed, harmonics, RMS amplitude? What physical and mathematical principles should I use to build these algorithms?
  • Obtaining ML training data: How will you obtain suitable datasets for ML training? In many cases this is not easy to obtain, as many foremen will not allow any disruption to their time-critical production lines.
  • Preparing datasets: After answering the aforementioned questions, the next challenge will be to capture and prepare the datasets for the ML classification. This is traditionally where a good 90% of a data scientist’s time will be spent. Therefore, it is prudent to invest in high fidelity feature extraction edge algorithms in order to expedite this step. This will also have the advantage of increasing the reproducibility and consistency of the results, which is where many AI based systems perform poorly.

ASN’s IP blocks and applications

ASN’s vibration IP blocks combine the Fourier transform’s time-frequency integration property, data filtering and a specialised high frequency resolution tracking algorithm to implement the ARAHTA (adaptive running speed and harmonics tracking) algorithm. ARAHTA tracks the vibration sensor’s ODR (output data rate) and calculates the motor/pumps running speed using the sensor’s accelerometer sensor data in real-time. ARAHTA’s high resolution and adaptive tracking mechanism results in a typical running speed accuracy of ±1 RPM across the temperature range and sub-mm displacement accuracy using noisy accelerometer data.

ARAHTA’s high accuracy and flexibility ensures that the resulting ML features are high quality and very consistent in the presence of temperature change and load shifts. This has a significant advantage for CbM applications, whereby fingerprinting a spectral profile can be used to assess the degradation of assets of interest. ARAHTA’s high-resolution spectrum forms the basis of providing an AI algorithm with high accuracy feature-rich information, suitable for classification.

Algorithmic performance

A comparison of the FFT vs the ASN ARAHTA IP blocks is shown below. Setting up a test accelerometer signal comprised of an 8.2Hz sinusoid with amplitude 1g and a few harmonic frequencies at various amplitudes, we can objectively compare the methods.

Analysing Figure 1, notice that the plot shows a comparison of the acceleration spectrum (i.e. the FFT of the acceleration data, shown in red) and the displacement spectrum, shown in blue. Analysing the first peak, notice that as the FFT’s resolution is insufficient, as the algorithm has identified the peak at 8.75Hz, rather than at 8.2Hz. This has a consequence for the amplitude estimation, as the acceleration spectrum amplitude is around 0.34g, rather than the expected 1g. As such, the algorithm incorrectly estimates the displacement at 8.2Hz to be 1mm, rather than 3.69mm.

The true value can be seen in Figure 2, where ARAHTA correctly finds the first resonant peak at 8.2Hz and estimates the correct amplitude of 3.69mm.

Figure 1 – Displacement estimate via FFT (frequency resolution: 813.5mHz):
wrong frequency and amplitude estimation
Figure 2 – Displacement estimate via ARAHTA (frequency resolution: 10mHz):
correct amplitude and frequency estimation.

Get in touch and reduce your asset’s TCO

ASN contactless measurement sensor technology and smart algorithms are an ideal solution for AI based CbM applications. Please contact our CbM expert team to see how we can help you create an effective maintenance programme and reduce your asset’s Total Cost of Ownership.

Author

  • Sanjeev is a RTEI (Real-Time Edge Intelligence) visionary and expert in signals and systems with a track record of successfully developing over 25 commercial products. He is a Distinguished Arm Ambassador and advises top international blue chip companies on their AIoT/RTEI solutions and strategies for I4.0, telemedicine, smart healthcare, smart grids and smart buildings.

    View all posts
/0 Comments/by

AIoT optimised DSP filtering library for Arm, RISC-V and MIPS microcontrollers

,

ASN Filter Designer’s new ANSI C SDK framework, provides developers with a comprehensive automatic C code generator for microcontrollers and embedded platforms. This allows developers to directly deploy their AIoT filtering application from within the tool to any STM32, Arduino, ESP32, PIC32, Beagle Bone and other Arm, RISC-V, MIPS microcontrollers for direct use.

Arm’s CMSIS-DSP library vs. ASN’s C SDK Framework

Thanks to our close collaboration with Arm’s architecture team, our new ultra-compact, highly optimised ANSI C based framework provides outstanding performance compared to other commercial DSP libraries, including Arm’s optimised CMSIS-DSP library.

Benchmarks for STM32: M3, M4F and M7F microcontrollers running an 8th order IIR biquad lowpass filter for 1024 samples

As seen, using o1 complier optimisation, our framework is able to surpass Arm’s CMSIS-DSP library’s performance on an M4F and M7F. Although notice that performance of both libraries is worse on the Cortex-M3, as it doesn’t have an FPU. Despite the difference, both libraries perform equally well, but the ASN DSP library has the added advantage of extra functionality and being platform agnostic, making it ideal for variety of biomedical (ECG, EMG, PPG), audio (sound effects, equalisers) , IoT (temperature, gas, pressure) and I4.0 (flow measurement, vibration analysis, CbM) applications.

AIoT applications designed on the newer Cortex-M33F and Cortex-M55F cores can also take advantage of extra filtering blocks, double precision arithmetic support, providing a simple way of implementing high performance AI on the Edge applications within hours.

Advantages for developers

  • A developer can now develop, test and deploy a complete DSP filtering application within the ASN Filter Designer within a few hours. This is very different from a traditional R&D approach that assigns a team of developers for several days in order to achieve the same level of accuracy required for the application.
  • Open source and agnostic code base: In order to allow developers to get the maximum performance for their applications, the ASN-DSP SDK is provided as open source and is written in ANSI C. This means that any embedded processor and any level of compiler optimisation can be used.
  • Memory size required for the ASN-DSP SDK is relativity lower than other standard DSP libraries, which makes the ASN-DSP SDK extremely suitable for microcontrollers that have memory constrains.
  • Using the ASN Filter Designer’s signal analyser tool, developers now can test the performance, accuracy and assess the frequency response of their designed filter and get optimised C code which they can directly use in their application.
  • The SDK also supports some extra filtering functions, such as: a median filter, a moving average filter, all-pass, single section IIR filters, a TKEO biomedical filter, and various non-linear functions, including RMS, Abs, Log and Sqrt.  These functions form the filter cascade within the tool, and can be used to build signal processing applications, such as EMG and ECG biomedical applications.
  • The ASN-DSP SDK supports both single and double precision floating point arithmetic, providing excellent numerical accuracy and wide dynamic range. The library is unique in the sense that it supports double precision arithmetic, which although is not the most optimal for microcontrollers, allows for the implementation of high-fidelity filtering applications.

The ANSI C SDK framework is further extended by our new C# .NET framework, allowing .NET developers to build high performance desktop applications with signal processing capabilities.

Find out more and try it yourself

Benchmarks on a variety of 32-bit embedded platforms, including a biomedical EMG filtering example, are covered in the following application note.

The both framework SDKs are available in ASNFD v5.0, which may be downloaded here.

ASN Filter Designer box
Download Demo
Pricing and Licencing

/0 Comments/by

Save time and energy in the design of the algorithms for Drones and DC motor control with ASN Filter Designer

,

Drones are one of the golden nuggets in AIoT. No wonder, they can play a pivotal role in congested cities and faraway areas for delivery. Further, they can be a great help to give an overview of a large area or for places which are difficult or dangerous to reach. Advanced Solutions did some research how the companies producing drones has solved some questions regarding their sensor technology. And in drones, there are a lot of sensors- and especial the DC motor control. We found out that with ASN Filter Designer, producers could have saved time and energy in the design of their algorithms with ASN Filter Designer.

Until now: hard-by found solutions

We found out that most producers had come very hard-by to their solutions. And that, when solutions are found, they are far from near perfect.

Probably, this producer has spent weeks or even months on finding these solutions. With ASN Filter Designer, he could have come to a solution within days or maybe hours. Besides, we expect that the measurement would be better too.

The most important issue is that algorithms were developed by handwork: developed in a ‘lab’ environment and then tried in real-life. With the result of the test, the algorithm would be adjusted again. Because a ‘lab’ environment where testing circumstances are stable, it’s very hard work to make the models work in ‘real’ life. For this, rounds and rounds of ‘lab development’ and ‘real life testing’ have to be made.

How ASN Filter Designer could have saved a lot of time and energy

ASN Filter Designer could have saved a lot of time in the design of the algorithms the following ways:

  • Design, analyze and implement filters for Drone senor applications 
  • Filters for speed and positioning control using sensorless BLDC motors
  • Speed up deployment

Real-time feedback and powerful signal analyzer

One of the key benefits of the ASN Filter Designer and signal analyzer is that it gives real-time feedback. Once an algorithm is developed, it can easily be tested on real-life data. To capture the real-life data, the ASN Filter Designer has a powerful signal analyzer in place. The tool’s signal analyzer implements a robust zero-crossings detector, allowing engineers to evaluate and fine-tune a complete sensorless BLDC control algorithm quickly and simply.

Design and analyze filters the easy way                         

You can easily design, analyze and implement filters for drone sensor applications. Including: loadcells, strain gauges, torque, pressure, temperature, vibration and ultrasonic sensors. And assess their dynamic performance in real-time with different input conditions.  With the ASN Filter Designer, no algorithms are needed: you just have to drag the filter design. The tool calculates the coordinates itself.

For speed and position control using sensorless BLDC (brushless DC) motors based on back-EMF filtering you can easily experiment with the ASN Filter Designer. See the results in real-time for various IIR, FIR and median (majority filtering) digital filtering schemes. The tool’s signal analyzer implements a robust zero-crossings detector. So you can evaluate and fine-tune a complete sensorless BLDC control algorithm quickly and simply.

Speed up deployment

Perform detailed time/frequency analysis on captured test datasets and fine-tune your design. Our Arm CMSIS-DSP and C/C++ code generators and software frameworks speed up deployment to a DSP, FPGA or micro-controller.

Drones use lots of sensors, and most challenges will be solved with them! ASN Filter Designer provides you with a simple way of improving your sensor measurement performance with its interactive design interface.

So, if you have a measurement problem, ask yourself: will I have a lot of frustrating and costs (maybe not ‘out of pocket’, but still: costs) of creating a filter by hand? Or would I create my filter within days or even hours and save a lot of headache and money. Because: it’s already possible to have a full 3-month license for only 140 euro!

ASN Filter Designer box
Download Demo
Pricing and Licencing
/0 Comments/by

Sensor market $87.6 billion by 2025

,

How do you get the best performance from your IoT smart sensor?

The global smart sensor market size is projected to grow from USD 36.6 billion in 2020 to USD 87.6 billion by 2025, at a CAGR of 19.0%. At least 80% of these IoT/IIoT smart sensors (temperature, pressure, gas, image, motion, loadcells) will use Arm’s Cortex-M technology.

IoT sensor measurement challenge

The challenge for most, is that many sensors used in these applications require filtering in order to clean the measurement data in order to make it useful for analysis.

Let’s have a look at what sensor data really is…. All sensors produce measurement data. These measurement data contain two types of components:

  • Wanted components, i.e. information what we want to know
  • Unwanted components, measurement noise, 50/60Hz powerline interference, glitches etc – what we don’t want to know

Unwanted components degrade system performance and need to be removed.

So, how do we do it?

DSP means Digital Signal Processing and is a mathematical recipe (algorithm) that can be applied to IoT sensor measurement data in order to clean it and make it useful for analysis.

But that’s not all! DSP algorithms can also help:

  • In analysing data, producing more accurate results for decision making with ML (machine learning)
  • They can also improve overall system performance with existing hardware. So ther’s no need to redesign your hardware: a massive cost saving!
  • To reduce the data sent off to the cloud by pre-analysing data. So send only the data which is necessary

Nevertheless, DSP has been considered by most to be a black art, limited only to those with a strong academic mathematical background. However, for many IoT/IIoT applications, DSP has been become a must in order to remain competitive and obtain high performance with relatively low cost hardware.

Do you have an example?

Consider the following application for gas sensor measurement (see the figure below). The requirement is to determine the amplitude of the sinusoid in order to get an estimate of gas concentration (bigger amplitude, more gas concentration etc). Analysing the figure, it is seen that the sinusoid is corrupted with measurement noise (shown in blue), and any estimate based on the blue signal will have a high degree of uncertainty about it – which is not very useful if getting an accurate reading of gas concentration!

Algorithms clean the sensor data

After ‘cleaning’ the sinusoid (red line) with a DSP filtering algorithm, we obtain a much more accurate and usable signal. Now we are able to estimate the amplitude/gas concentration. Notice how easy it is to determine the amplitude of red line.

This is only a snippet of what is possible with DSP algorithms for IoT/IIoT applications, but it should give you a good idea as to the possibilities of DSP.

How do I use this in my IoT application?

As mentioned at the beginning of this article, 80% of IoT smart sensor devices are deployed on Arm’s Cortex-M technology. The Arm Cortex-M4 is a very popular choice with hundreds of silicon vendors, as it offers DSP functionality traditionally found in more expensive DSPs. Arm and its partners provide developers with easy to use tooling and a free software framework (CMSIS-DSP). So, you’ll be up and running within minutes.

/0 Comments/by

Difference between IIR and FIR filters: a practical design guide

,

A digital filter is a mathematical algorithm that operates on a digital dataset (e.g. sensor data) in order extract information of interest and remove any unwanted information. Applications of this type of technology, include removing glitches from sensor data or even cleaning up noise on a measured signal for easier data analysis. But how do we choose the best type of digital filter for our application? And what are the differences between an IIR filter and an FIR filter?

Digital filters are divided into the following two categories:

  • Infinite impulse response (IIR)
  • Finite impulse response (FIR)

As the names suggest, each type of filter is categorised by the length of its impulse response. However, before beginning with a detailed mathematical analysis, it is prudent to appreciate the differences in performance and characteristics of each type of filter.

Example

In order to illustrate the differences between an IIR and FIR, the frequency response of a 14th order FIR (solid line), and a 4th order Chebyshev Type I IIR (dashed line) is shown below in Figure 1.  Notice that although the magnitude spectra have a similar degree of attenuation, the phase spectrum of the IIR filter is non-linear in the passband (\(\small 0\rightarrow7.5Hz\)), and becomes very non-linear at the cut-off frequency, \(\small f_c=7.5Hz\). Also notice that the FIR requires a higher number of coefficients (15 vs the IIR’s 10) to match the attenuation characteristics of the IIR.

FIR vs IIR: frequency response of a 14th order FIR (solid line), and a 4th order Chebyshev Type I IIR (dashed line); Fir Filter, IIR Filter
Figure 1: FIR vs IIR: frequency response of a 14th order FIR (solid line), and a 4th order Chebyshev Type I IIR (dashed line)

These are just some of the differences between the two types of filters. A detailed summary of the main advantages and disadvantages of each type of filter will now follow.

IIR filters

IIR (infinite impulse response) filters are generally chosen for applications where linear phase is not too important and memory is limited. They have been widely deployed in audio equalisation, biomedical sensor signal processing, IoT/IIoT smart sensors and high-speed telecommunication/RF applications.

Advantages

  • Low implementation cost: requires less coefficients and memory than FIR filters in order to satisfy a similar set of specifications, i.e., cut-off frequency and stopband attenuation.
  • Low latency: suitable for real-time control and very high-speed RF applications by virtue of the low number of coefficients.
  • Analog equivalent: May be used for mimicking the characteristics of analog filters using s-z plane mapping transforms.

Disadvantages

  • Non-linear phase characteristics: The phase charactersitics of an IIR filter are generally nonlinear, especially near the cut-off frequencies. All-pass equalisation filters can be used in order to improve the passband phase characteristics.
  • More detailed analysis: Requires more scaling and numeric overflow analysis when implemented in fixed point. The Direct form II filter structure is especially sensitive to the effects of quantisation, and requires special care during the design phase.
  • Numerical stability: Less numerically stable than their FIR (finite impulse response) counterparts, due to the feedback paths.

FIR filters

FIR (finite impulse response) filters are generally chosen for applications where linear phase is important and a decent amount of memory and computational performance are available. They have a widely deployed in audio and biomedical signal enhancement applications. Their all-zero structure (discussed below) ensures that they never become unstable for any type of input signal, which gives them a distinct advantage over the IIR.

Advantages

  • Linear phase: FIRs can be easily designed to have linear phase. This means that no phase distortion is introduced into the signal to be filtered, as all frequencies are shifted in time by the same amount – thus maintaining their relative harmonic relationships (i.e. constant group and phase delay). This is certainly not case with IIR filters, that have a non-linear phase characteristic.   
  • Stability: As FIRs do not use previous output values to compute their present output, i.e. they have no feedback, they can never become unstable for any type of input signal, which is gives them a distinct advantage over IIR filters.
  • Arbitrary frequency response: The Parks-McClellan and ASN FilterScript’s firarb() function allow for the design of an FIR with an arbitrary magnitude response. This means that an FIR can be customised more easily than an IIR.
  • Fixed point performance: the effects of quantisation are less severe than that of an IIR.

Disadvantages

  • High computational and memory requirement: FIRs usually require many more coefficients for achieving a sharp cut-off than their IIR counterparts. The consequence of this is that they require much more memory and significantly a higher amount of MAC (multiple and accumulate) operations. However, modern microcontroller architectures based on the Arm’s Cortex-M cores now include DSP hardware support via SIMD (signal instruction, multiple data) that expedite the filtering operation significantly.
  • Higher latency: the higher number of coefficients, means that in general a linear phase FIR is less suitable than an IIR for fast high throughput applications. This becomes problematic for real-time closed-loop control applications, where a linear phase FIR filter may have too much group delay to achieve loop stability.
  • Minimum phase filters: A solution to ovecome the inherent N/2 latency (group delay) in a linear filter is to use a so-called minimum phase filter, whereby any zeros outside of the unit circle are moved to their conjugate reciprocal locations inside the unit circle. The result of the zero flipping operation is that the magnitude spectrum will be identical to the original filter, and the phase will be nonlinear, but most importantly the latency will be reduced from N/2 to something much smaller (although non-constant), making it suitable for real-time control applications.
          For applications where phase is less important, this may sound ideal, but the difficulty arises in the numerical accuracy of the root-finding algorithm when dealing with large polynomials. Therefore, orders of 50 or 60 should be considered a maximum when using this approach. Although other methods do exist (e.g. the Complex Cepstrum), transforming higher-order linear phase FIRs to their minimum phase cousins remains a challenging task.
  • No analog equivalent: using the Bilinear, matched z-transform (s-z mapping), an analog filter can be easily be transformed into an equivalent IIR filter.  However, this is not possible for an FIR as it has no analog equivalent.

Mathematical definitions

As discussed in the introduction, the name IIR and FIR originate from the mathematical definitions of each type of filter, i.e. an IIR filter is categorised by its theoretically infinite impulse response,

\(\displaystyle
y(n)=\sum_{k=0}^{\infty}h(k)x(n-k)
\)

and an FIR categorised by its finite impulse response,

\(\displaystyle
y(n)=\sum_{k=0}^{N-1}h(k)x(n-k)
\)

We will now analyse the mathematical properties of each type of filter in turn.

IIR definition

As seen above, an IIR filter is categorised by its theoretically infinite impulse response,

\(\displaystyle y(n)=\sum_{k=0}^{\infty}h(k)x(n-k) \)

Practically speaking, it is not possible to compute the output of an IIR using this equation. Therefore, the equation may be re-written in terms of a finite number of poles \(\small p\) and zeros \(\small q\), as defined by the linear constant coefficient difference equation given by:

\(\displaystyle
y(n)=\sum_{k=0}^{q}b_k x(n-k)-\sum_{k=1}^{p}a_ky(n-k)
\)

where, \(\small a_k\) and \(\small b_k\) are the filter’s denominator and numerator polynomial coefficients, who’s roots are equal to the filter’s poles and zeros respectively. Thus, a relationship between the difference equation and the z-transform (transfer function) may therefore be defined by using the z-transform delay property such that,

\(\displaystyle
\sum_{k=0}^{q}b_kx(n-k)-\sum_{k=1}^{p}a_ky(n-k)\quad\stackrel{\displaystyle\mathcal{Z}}{\longleftrightarrow}\quad\frac{\sum\limits_{k=0}^q b_kz^{-k}}{1+\sum\limits_{k=1}^p a_kz^{-k}}
\)

As seen, the transfer function is a frequency domain representation of the filter. Notice also that the poles act on the output data, and the zeros on the input data. Since the poles act on the output data, and affect stability, it is essential that their radii remain inside the unit circle (i.e. <1) for BIBO (bounded input, bounded output) stability. The radii of the zeros are less critical, as they do not affect filter stability. This is the primary reason why all-zero FIR (finite impulse response) filters are always stable.

BIBO stability

A linear time invariant (LTI) system (such as a digital filter) is said to be bounded input, bounded output stable, or BIBO stable, if every bounded input gives rise to a bounded output, as

\(\displaystyle \sum_{k=0}^{\infty}\left|h(k)\right|<\infty \)

Where, \(\small h(k)\) is the LTI system’s impulse response. Analyzing this equation, it should be clear that the BIBO stability criterion will only be satisfied if the system’s poles lie inside the unit circle, since the system’s ROC (region of convergence) must include the unit circle. Consequently, it is sufficient to say that a bounded input signal will always produce a bounded output signal if all the poles lie inside the unit circle.

The zeros on the other hand, are not constrained by this requirement, and as a consequence may lie anywhere on z-plane, since they do not directly affect system stability. Therefore, a system stability analysis may be undertaken by firstly calculating the roots of the transfer function (i.e., roots of the numerator and denominator polynomials) and then plotting the corresponding poles and zeros upon the z-plane.

An interesting situation arises if any poles lie on the unit circle, since the system is said to be marginally stable, as it is neither stable or unstable. Although marginally stable systems are not BIBO stable, they have been exploited by digital oscillator designers, since their impulse response provides a simple method of generating sine waves, which have proved to be invaluable in the field of telecommunications.

Biquad IIR filters

The IIR filter implementation discussed herein is said to be biquad, since it has two poles and two zeros as illustrated below in Figure 2. The biquad implementation is particularly useful for fixed point implementations, as the effects of quantization and numerical stability are minimised. However, the overall success of any biquad implementation is dependent upon the available number precision, which must be sufficient enough in order to ensure that the quantised poles are always inside the unit circle.

Direct Form I (biquad) IIR filter realization and transfer function.; Direct Form; Biquad filter

Figure 2: Direct Form I (biquad) IIR filter realization and transfer function.

Analysing Figure 2, it can be seen that the biquad structure is actually comprised of two feedback paths (scaled by \(\small a_1\) and \(\small a_2\)), three feed forward paths (scaled by \(\small b_0, b_1\) and \(\small b_2\)) and a section gain, \(\small K\). Thus, the filtering operation of Figure 1 can be summarised by the following simple recursive equation:

\(\displaystyle y(n)=K\times\Big[b_0 x(n) + b_1 x(n-1) + b_2 x(n-2)\Big] – a_1 y(n-1)-a_2 y(n-2)\)

Analysing the equation, notice that the biquad implementation only requires four additions (requiring only one accumulator) and five multiplications, which can be easily accommodated on any Cortex-M microcontroller. The section gain, \(\small K\) may also be pre-multiplied with the forward path coefficients before implementation.

A collection of Biquad filters is referred to as a Biquad Cascade, as illustrated below.

Biquad Cascade; Biquad filter

The ASN Filter Designer can design and implement a cascade of up to 50 biquads (Professional edition only).

Floating point implementation

When implementing a filter in floating point (i.e. using double or single precision arithmetic) Direct Form II structures are considered to be a better choice than the Direct Form I structure. The Direct Form II Transposed structure is considered the most numerically accurate for floating point implementation, as the undesirable effects of numerical swamping are minimised as seen by analysing the difference equations.

Direct Form II Transposed strucutre, transfer function and difference equations; IIR Filters; Biquad Filters

Figure 3 – Direct Form II Transposed strucutre, transfer function and difference equations

The filter summary (shown in Figure 4) provides the designer with a detailed overview of the designed filter, including a detailed summary of the technical specifications and the filter coefficients, which presents a quick and simple route to documenting your design.

The ASN Filter Designer supports the design and implementation of both single section and Biquad (default setting) IIR filters.

Biquad filter ASN Filter Designer DSP
Figure 4: detailed specification.

FIR definition

Returning the IIR’s linear constant coefficient difference equation, i.e.

\(\displaystyle
y(n)=\sum_{k=0}^{q}b_kx(n-k)-\sum_{k=1}^{p}a_ky(n-k)
\)

Notice that when we set the \(\small a_k\) coefficients (i.e. the feedback) to zero, the definition reduces to our original the FIR filter definition, meaning that the FIR computation is just based on past and present inputs values, namely:

\(\displaystyle
y(n)=\sum_{k=0}^{q}b_kx(n-k)
\)

Implementation

Although several practical implementations for FIRs exist, the direct form structure and its transposed cousin are perhaps the most commonly used, and as such, all designed filter coefficients are intended for implementation in a Direct form structure.

The Direct form structure and associated difference equation are shown below. The Direct Form is advocated for fixed point implementation by virtue of the single accumulator concept.

\(\displaystyle y(n) = b_0x(n) + b_1x(n-1) + b_2x(n-2) + …. +b_qx(n-q) \)

Direct form; Direct form structure

The recommended (default) structure within the ASN Filter Designer is the Direct Form Transposed structure, as this offers superior numerical accuracy when using floating point arithmetic. This can be readily seen by analysing the difference equations below (used for implementation), as the undesirable effects of numerical swamping are minimised, since floating point addition is performed on numbers of similar magnitude.

\(\displaystyle \begin{eqnarray}y(n) & = &b_0x(n) &+& w_1(n-1) \\ w_1(n)&=&b_1x(n) &+& w_2(n-1) \\ w_2(n)&=&b_2x(n) &+& w_3(n-1) \\ \vdots\quad &=& \quad\vdots &+&\quad\vdots \\ w_q(n)&=&b_qx(n) \end{eqnarray}\)

Direct form Transposed

Implementing your filter on an Arm Cortex-M processor

Although a few processor technologies exist for microcontrollers (e.g. RISC-V, Xtensa, MIPS), over 90% of the microcontrollers used in the smart product market are powered by so-called Arm Cortex-M processors that offer a combination of high algorithmic performance, low-power and security. The Arm Cortex-M4 is a very popular choice with several silicon vendors (including ST, TI, NXP, ADI, Nordic, Microchip, Renesas), as it offers DSP (digital signal processing) functionality traditionally found in more expensive devices and is low-power.

Filtering libraries and support

Arm and ASN provide developers with extensive easy-to-use tooling and tried and tested software libraries used internationally by tens of thousands of developers.

The Arm CMSIS-DSP software framework is interesting as it provides IoT developers with a rich collection of fast mathematical and vector functions, interpolation functions, digital filtering (FIR/IIR) and adaptive filtering (LMS) functions, motor control functions (e.g. PID controller), complex math functions and supports various data types, including fixed and floating point. The important point to make here is that all of these functions have been optimised for Arm Cortex-M processors, allowing you to focus on your application rather than worrying about optimisation.  

Despite the broad functionality, the CMSIS-DSP library is somewhat limited for filters, so the flexible ASN DSP filtering library can be used instead, which supports the higher numerical accuracy Direct Form Transposed FIR filter structure and single section IIR filters. A benchmark of ASN’s floating point application-specific DSP filtering library versus Arm’s CMSIS-DSP library is shown below for three types of Arm cores.

Framework Benchmarks: lower number of clock cycles means higher performance.

As seen, the performance of the ASN library is slightly faster by virtue of the application-specific nature of the implementation. The C code is automatically generated from the ASN Filter Designer tool.

What have we learned?

Digital filters are divided into the following two categories:

  • Infinite impulse response (IIR)
  • Finite impulse response (FIR)

IIR (infinite impulse response) filters are generally chosen for applications where linear phase is not too important and memory is limited. They have been widely deployed in audio equalisation, biomedical sensor signal processing, IoT/AIoT smart sensors and high-speed telecommunication/RF applications.

FIR (finite impulse response) filters are generally chosen for applications where linear phase is important and a decent amount of memory and computational performance are available. They have a widely deployed in audio and biomedical signal enhancement applications.

ASN Filter Designer provides engineers with everything they need to design, experiment and deploy complex IIR and FIR digital filters for a variety of IoT sensor measurement applications. These advantages coupled with automatic C code generation with ASN’s DSP filtering library functionality allow engineers to design, validate and then deploy their designs to an Arm Cortex-M processor within hours rather than more traditional routes that could take days.

 

 

Download demo now

Licencing information

Author

  • Sanjeev is a RTEI (Real-Time Edge Intelligence) visionary and expert in signals and systems with a track record of successfully developing over 25 commercial products. He is a Distinguished Arm Ambassador and advises top international blue chip companies on their AIoT/RTEI solutions and strategies for I4.0, telemedicine, smart healthcare, smart grids and smart buildings.

    View all posts
/by

Covid-19: Netherlands and Germany death rates stabilise, but Belgium still skyrockets

,

Continuing with our Analytics team study of the virus on Western European countries, we present our findings for data up to week 15 (14 April).

As discussed in our previous articles, in order to provide an objective comparison per country, the algorithmic results need to be standardised around the population of each country in order to produce a more accurate deaths per million inhabitants rate. The figure shown below summarises the results.

As seen, Belgium’s mortality rate (red) is significantly higher than any of its neighbours. Germany (blue) and the Netherlands (green) have the lowest mortality rates, and appear to be levelling off. This suggests that the Dutch and German governments testing, health care systems and social distancing strategies appear to be paying off.

It’s not completely clear why Belgium’s mortality rate is so much higher than its neighbours, but a possible explanation may be due to insufficient testing and the virus hitting various elderly care homes. We’ll follow Belgium’s progress over the coming weeks, and report our findings.

The UK

As discussed in a previous article, the UK had a one-week head start on its neighbours. Therefore, shifting the UK data left by six days, we obtain an interesting picture of the UK’s situation:

Applying a prediction model to the UK data (dashed magenta line), notice how the UK’s data follows France’s data. Although long term predication models should be viewed with a degree of scepticism (as there are too many unknown factors to consider), the prediction suggests that the UK’s mortality rate should follow France’s mortality rate.  

The good news for the UK population, is that the emergency measures in place, appear to be working and are leading to a decline in deaths!

/0 Comments/by

6 Strengths of the ASN Filter Designer DSP platform

,

6 reasons why ASN Filter Designer is a powerful real-time DSP platform e.g. life math scripting, tool creates your technical specification and documentation

Read more
/0 Comments/by

Covid-19: UK had 1 week headstart, but failed to act

,

The Covid-19 virus has forced European governments to order millions to lockdown in the hope of limiting the spread of the virus, based on ‘expert scientific advice’.  The latest recent review of WHO data by Dutch data modelling specialist, Advanced Solutions Nederland (ASN) reveals that the UK could of adverted strain on services and avoided a sharp rise in Covid-19 cases by taking advantage of being six days behind the infection spread in Northern  Europe, but failed to put measures in place in time, due to flawed ‘expert’ predictions.

Central to government policies imposed are predictions being made from statistics that are essentially handling raw data ineffectively. Many models are based on raw measured values that are not adjusted for comparison with neighbouring countries, so called population standardisation, which can give a false perspective of the situation at hand.

– Director of Algorithms and Analytics, ASN, Dr. Sanjeev Sarpal

Ineffective use of modelling to predict virus trend

John Hopkins University (JHU) provide an open database of confirmed cases, deaths and number of recoveries, obtained from data from the World Health Organisation (WHO), various other health intuitions and governments. These datasets are broken down into countries and regions.

Analysis considered data obtained from the following five European countries populations: Germany (83 million), France (67 million) UK (66 million), the Netherlands (17 million), Belgium (11 million).

Our analysts found that by analysing the viral trend by doing a ‘like with like’ comparison of populations rather than the conventional method of non-standardisation, resulted in a totally contradicting set of results, implying that the UK governments response was not informed appropriately.

In order to provide an objective comparison per country, the algorithmics results were standardised around the population of each country in order to produce a more accurate deaths per million inhabitants rate. The figure shown below summarises the results.

Analysing the chart, it can be seen that all central countries considered herein all report first cases within days of each other, and have very similar contamination rate. The UK is the exception, as it is approximately 6 days behind mainland Europe.

By shifting the UK left by six days, we see that the UK also follows the same trend as its continental neighbours. The dashed line represents the algorithmic prediction of the number of confirmed cases for the next two days (short term prediction), which closely follows the other countries.

Thus, it can be concluded that despite the British government having advanced warning, they failed to adequately prepare themselves for the effects of the virus.

No magic long-term prediction model

There are a multitude of data modelling methods, each giving a different result depending on the interpretation required. For the Covid-19 virus, there is no ‘magic model’ that can be used to predict the long-term severity of the outbreak, as there are too many variables to consider, which are almost impossible to model and track as the pandemic unfolds.

External factors, such as emergency laws, increased public hygiene/diligence and better medical care facilities are but a few major factors that affect any long-term prediction model. These critical factors are generally not modelled when making a prediction model. The short-term prediction shown herein, was just for the next two days, but all prediction models must be viewed with a degree of scepticism, as it is not possible to model all of the unique circumstances that present themselves.  

ASN’s data analytics team will be closely monitoring the development of the Covid-19 virus, and providing regular updates via our blog.

/0 Comments/by

Power Blackouts: jargon buster

,

Energy companies have struggled for years with meeting demand with supply with society’s increasing demand for energy. This been made even more challenging with more people using electric vehicles and smart cities demanding more lighting.

Modern IoT sensors and smart grid solutions help energy companies and consumers improve and optimize the modern grid for the 21st century. But what does all the jargon really mean?

Blackout

The UK National Grid recently experienced a major outage that left almost a million homes in the dark and forced trains to a standstill. The source of the blackout was traced back to two generators that failed, resulting in grid’s frequency falling below the critical 49.5Hz set by the regulator.

According to the media the UK blackout was triggered when the frequency slumped to 48.88Hz, which is well below the legal limits set by the regulatory agencies.

But what do these limits really mean?

Some background information

The energy grid frequency is 50Hz in Europe, 60Hz in the US. Japan has an unusual historical situation in that the East of the country runs on a European 50Hz system and the West of country runs on an American 60Hz system.

In all cases, in order to meet the energy requirements, several generators are needed to work in parallel and must be synchronised. Accurate frequency control is required to control the amount of power delivered by multiple generators in order to provide a stable power supply to consumers. The challenge for the energy companies is meeting the changes in supply and demand, since higher demand than supply will result in fall of frequency and vice versa.

Thus, the challenge for IoT sensors and algorithms is measuring the operating frequency and phase to a sufficient accuracy and adjusting the generators to meet the energy demand requirement at that particular time. But how?

A PMU (phase measurement unit) is typically used the measure and report back (typically 30-60 measurements per second) to the network operator what the actual frequency and phase of various points on the grid are. In order to synchronise the measurements, the PMU internal clocks are time synchronised via a GPS (global positioning system) unit, such that all reported frequency and phase measured across the grid are time aligned.

The frequency limits are shown below:

The challenge for energy managers

As seen above, the normal region in Europe is between 49.85 – 50.15Hz. If the generators exceed 50.15Hz (entering the orange region), there is too much energy and the generators need to be rolled back a little. If the frequency falls below 49.85Hz (also in the orange region), there is not enough energy to meet demand, and more energy is needed. In all cases, the frequency must never enter the red region, otherwise Blackouts will occur.

The energy company is legally obliged to keep the powerline frequency between 49.5 – 50.5Hz (± 1%). This is typically tracked to an accuracy of ± 1mHz resolution.

Blackouts

The UK blackout was triggered when the frequency slumped to 48.88Hz, which is well below the legal limits and in the blackout region. The damage to the UK economy has still yet to be determined, but National Grid UK should be considering adding extra redundancy safe guards in order restore public confidence.

Dips and swells tracking

Another common problem that occurs is that of energy dips, i.e. the voltage momentarily drops for a few cycles. Think about lights temporarily flickering in your house.

In factories running machinery, this usually occurs when a machine is started up, indicating imminent component failure. Swells are the opposite of dips, but are much less common.

ASN’s IoT sensor and algorithms play an essential role in keeping the grid healthy, as demonstrated in the video below.

/0 Comments/by