EXTENSOMETRIC HYDRODYNAMIC TRANSDUCER

FOR INFLUENCE SEA-MINE FUSES

 SENSOR CONSTRUCTION AND DEVELOPMENT OF SIGNAL PROCESSING ALGORITHMS FOR HYDRODYNAMIC FIELD ANALYSIS 

            Fuses of modern influence mines are usually equipped with several measurement channels like: acoustic, magnetic, electric, seismic and of course hydrodynamic. Most of these channels can be simulated by modern sweeps quite accurately, however the hydrodynamic channel is the one that can not be simulated in any accurate way.

            The lack of good simulators (sweeps) makes hydrodynamic channel a natural choice for a main mine activation method, especially in shallow water. However, as experience shows, the measurement conditions for a ship's hydrodynamic field are most of the time less then ideal. Sea waves, water currents or high tides can severely distort sensor output, leading to errors in ships detection, or even worse to false activation of the mine. There are also other constrains that need to be taken into account when designing and building hydrodynamic sensor for sea mines like: power supply, working depths, required sensitivity and other.

            In OBR CTM a new hydrodynamic sensor designed especially for bottom and anchor sea mines have been developed and it is now being actively tested at the sea and in the laboratory. As well as the sensor electronics, a new signal processing algorithms are developed to improve detection level in unideal environmental conditions or during high tides.

Hydrodynamic field generated by ships

            Every moving ship generate hydrodynamic filed (pressure changes) that can be detected by the mines sensor. This field depends on ship velocity, ship size, hull submersion and others. What will be detected by sensor also depends on current sea conditions, and sometimes these conditions can have major impact on acquired data.                   

            In ideal conditions ship's hull and even waves after stern are clearly visible and also relatively easy to detect, even by simply looking at field waveforms. In such conditions almost every sensor with required sensitivity and very simple algorithms can be used.

            Unfortunately sea level 0 occurs very rarely, and it is necessary that during development process, the transducer is prepared to work in different conditions (for example high waves). For sea trials performed with about 1 meter high, very irregular waves (sea level 3), that are quite typical for Baltic sea near harbours the ship passing over the mine was almost impossible to detect using only hydrodynamic sensor's data. In fact acoustic “peak” (recorded by another sensor) was the only way to do so. In such conditions no simple detection algorithm can be used to accurately detect the ship with only hydrodynamic channel data. Irregular waves and possibly also underwater currents causes signals that are comparable or even higher in amplitude then the signal generated by the ship.

Requirements for the hydrodynamic sensor.

            Using acquired data and NAVY's requirements a set of characteristics that the sensor should fulfil have been prepared. Among others the most important are:

Sensor has to work with static depths (mine operating depths) up to 100 meters,

Sensor has to be powered from mine internal battery and should have very low power consumption,

Sensor has to be turned on and off by mine main control processor and should be ready to work within 2-3 seconds after turn-on,

Sensor must measure small pressure changes from DC to a few Hz, with sensitivity that is constant over all working depths,

Sensitivity should be not worse then 1Pa (0.1 mm of water) and preferably better.

            Sea mines are usually prepared for very long working time and they have to be fully autonomic, so capacitance of power supply is large limiting factor for developers. The problem is normally solved by equipping mines with several “stand-by” and “combat” circuits, that works together. Stand-by circuits are designed to have minimum possible power consumption and their only task is to detect coming ship (or signs of ship) and turn on combat circuits that perform accurate detection and activation of mine. Combat circuits are most of the time turned off and do not consume power, but they still have to designed with power efficiency in mind.

            Almost all algorithms that are used to detect the ship works best when accurate data closely following real pressure changes that are presented on sensor input. This means that sensor should not introduce any filtering on the signal (high-pass behaviour is mostly problematic), and should respond irrespective of working depth. Sensing of very low frequencies is needed to detect slowly moving ships, while the pass band up to a few Hz is normally more then enough.

            Large working depth range (up to 100m) that have been requested, in connection with required sensitivity is almost impossible to meet using standard measurement technique because, dynamic range would be to high for low power circuits. One of the possible solutions is a compensation circuit that bring sensor output to 0 at any given depth. This circuit can also be used to measure the current working depth of the mine.

Development and construction of hydrodynamic sensor with active static pressure compensation circuit.

            After long research we decided that the extensometric absolute pressure transducers are the most suitable for the hydrodynamic sensor. They have low power consumption, are highly linear and it is possible to build a sensor that has no moving (mechanical) parts. Mechanical parts are known source of reliability problems especially after they are submerged in sea water for a long time. Using two transducers instead of one increases sensitivity, reduces noise floor and allows using differential electronic circuits to process the transducer's output.

            As already have been told it is impossible to build low power circuit that would have enough dynamic range to reach required sensitivity with so large static depths changes. As a solution a special compensation circuit have been developed. The compensation circuit consists of resistors and variable resistors network controlled from sensor's main micro-controller.

            The micro-controller uses variable resistors to bring sensors outputs to 0 at any given working depth. The speed of the compensation procedure is limited mostly by low-pass signal filtering (to clean unwanted high frequency noise) of the amplifiers. To reduce compensation time to absolute minimum an idea from SAR (Successive Approximation Register) A/D converters have been used – micro controller tests all of the bits in binary compensation word starting with the most important and going to the least important bit. Using this technique and 10 bit compensation word we can achieve less then 1 second compensation time and also, by using the compensation results the current depth can be measured with a few centimetres accuracy.

            After construction a lot of laboratory tests have been carried out on the sensor to verify it's resolution, depths range, power consumption and others. A special stand have been build to test compensation circuits and resolution on various static depths as well as compensation range and depth measurement accuracy. The noise floor have been tested using another stand, but it turns out that electronics and transducers noise have to tested separately. Normal pressure variations in a laboratory (outside winds, closing or opening windows or even floor vibration) are within sensors detection range and can distorts the results.

            During testing, the sensor was submitted to a various static pressures from 100kPa (normal atmospheric pressure) to 1.1MPa (100 meter of water). Then, after compensation procedure has been finished, a very small (100 to 5000Pa) dynamic pressure variations was administered using a special pump, with controlled rate of change, value and shape. During all of the time sensor output was recorded and reviewed. These measurements shown that our sensor successfully fulfilled the requirements, while its resolution and accuracy of depth measurement is even better then required. Sensor output is directly proportional to input pressure without any unwanted filtration or disturbances, that simplifies design of signal processing algorithms for ship detection.

Signal processing algorithms.

            Ship's hydrodynamic field can be severely distorted by sea conditions, especially by waves and underwater currents, as already was shown. To improve ship detection probability a special signal processing algorithms have to be used, that makes use of adaptive filtering and other advanced signal processing methods. Generally all these methods tries to “learn” how the noise signal looks like and then detect disturbance that looks “different” - such disturbance might be a ship's field that we are looking for. Such filtration is also very effective in dealing with slow constant depth changes (high tides for example) as well as with constant frequency false signals. In both cases filter removes the unwanted signals after a few seconds of learning.              

Conclusion

Hydrodynamic field generated by the ship can be heavily distorted by sea waves or underwater currents,

To separate signals caused by waves and other disturbances, a special signal processing algorithms was elaborated, that bases on adaptive filters,

Presented sensor (with two extensometric transducers), that incorporates no moving parts can quickly compensate constant static pressure, irrespective of working depth (up to 100m) using algorithm similar to SAR analogue to digital converters,

Static pressure compensation allows sensor to measure small hydrodynamic fields at large working depths (large static pressures), without high dynamic electronics circuits (about 22 bits would be required) that normally have too high power consumption,

Sensor can precisely reproduce hydrodynamic pressure changes irrespective of signal frequency, shape and static depth. Precise reproduction of hydrodynamic pressure is required for proper operation of used signal processing algorithms.

Developed sensor is scheduled for sea trials with Navy ships, to test its parameters measured in the laboratory and to tune signal processing algorithms.

Acknowledgement:

            This work is financed by found dedicated for research and development projects realized during the years 2009-2011:  Project Number 399/BO/A.

more details: Widlok 11_02.pdf 63.1 KB