With the development of computer software technology, people can imagine the flaw of weighing sensor itself through software technology to solve? That is to say, by the computer software to complete sensors such as zero compensation, temperature compensation, and linear compensation creep and recovery, lag compensation, compensation, etc. Almost all of the compensation process. This can make the making craft of weighing sensor itself becomes extremely simple, does not need to spend a lot of energy on the fine craftsmanship, and can greatly improve the percent of pass of the weighing sensor elastomer with the patch. Of course, due to complete the above all kinds of software compensation, need various mathematical model is set up, need large database to support. Not a lot of test data is impossible.

The basic configuration is as follows:

Analog sensor + digital analog (amplification and A/D circuit) + sensor intelligent software compensation

The weighing sensor of digital analog part includes amplification, filtering, A/D converter, microprocessors, temperature sensor, through the digital compensation circuit and digital compensation technology, linearity, hysteresis and creep can be compensated; Built-in temperature sensor, through the software compensation to the real-time temperature compensation; Address is adjustable, convenient for application and interchangeable; And can realize remote diagnosis and correction. Its core is intelligent sensor software compensation technology. The weighing sensor is said to have used the theory of fuzzy mathematics, artificial intelligence, etc, with the data processing method for the digital compensation of sensor error, avoid the traditional analog compensation method of weighing sensor. This kind of sensor has the basic requirements of intelligent digital compensation technology.

A kind of using neural network self-learning function, solved due to the change of environmental temperature on the weighing sensor characteristic differences between bridge arm measurement error caused by the impact. Particular way is: to bridge the two output voltage signals as the calibration data, using neural network data fusion of the calibration data for processing, thus improving both the suitable temperature range of bridge, and improve its static characteristic.

At present, the domestic weighing sensor nonlinear rely mainly on the elastomer itself manufacture, compensation, adjust the process to solve. And a way to use BP algorithm with the nonlinear mapping ability of software for sensor calibration data input - output characteristic of nonlinear approximation, nonlinear correction as the intelligent sensor system software, the sensor with the support of the software to improve measurement accuracy. Sensor experimental data by BP neural network, according to this method can decrease the measuring relative error.

The smart sensor data pretreatment method, applied to the nonlinear correction of sensor temperature compensation, digital filter and scale transformation, which can realize industrial field sensor test data of the front-end detection and treatment, so as to improve the automation testing operation quality of the sensor in the system.