For more information on National Instruments hardware and LabVIEW please visit there website below

Data Acquisition (DAQ) is the automatic collection of data from sensors, instruments and devices on the factory floor, in the laboratory or the field. 

DAQ allows the interpretation of Thermocouple, Current or Voltage signals from any analog or digital transducer. Rothenberg Industries, LLC uses this technology to design turnkey process control solutions.

A common scenario is controlling the temperature of an object by reading the voltage signals generates by thermocouple(s) located in the area of interest. A LabVIEW interprets those signals via the DAQ device and compares the temperature to the set-point value, adjusting a output signal accordingly to the heat controller. 

case studies

Here are some examples of processes that Rothenberg Industries, LLC has develped

LabVIEW Ethernet Enabled Manufacturing Data Acquisition System

Mechanical Engineering six-sigma, mechanical-engineering, manufacturing, excel-vba, labview

Our client has dozens of manufacturing furnaces and hydraulic presses that require data acquisition with traceability down to the individual component serial number. The system would be interfaced from multiple shop floor computers around the plant and from a central engineering panel. Our client required that their data acquired for thermocouple, current and voltage transducer signals be easily accessible. The process was to be designed for future scalability regarding operator stations, data acquisition channels and devices. We developed a LabVIEW data acquisition system that continuously runs on a protected networked terminal server that controls multiple ethernet enabled data acquisition devices. A qualified engineer can remote into the terminal server from any location to interface with the central engineering panel. During production, the system is interfaced by one or more operator terminals that are run off shop floor computers. The system is ergonomically configured from a Microsoft excel file including data acquisition, operator panel, real-time alarm, and post-run KPI settings. The data acquisition settings include the Static IP of the ethernet device(s), the individual DAQ modules, the individual channels in the module and signal type (Thermocouple type, Current or Voltage). The sampling rate can be individually set for each channel as well as the data acquisition start and end value. If the channel type to be acquired is a current or voltage signal, a linear scaling factor is available. The system also includes several real-time alarms such as over-temp and uniformity warnings that are displayed on the corresponding operator panels set from the Microsoft excel configuration file. The data is continuously written to a *.flat file in real time and stored on a segregated terminal server to preserve partial data if there is any network interruption. Upon completion of a run, a VBA script is triggered to analyze the raw data and calculate predefined cycle metrics that are stored a manufacturing SQL database along with the respective production information. The production metrics are to be used for statistical process control via JMP JSL.

Operator panel communication

Flow Diagram - Ethernet enabled data acquisition system

Engineering terminal server interface

Operator shop-floor interface

Numerical Simulations

Tank Level PID Controller Simulation using the LabVIEW Control and Simulation Module with Auto-tuning & Gain Scheduling

Mechanical Engineering six-sigma, manufacturing, labview, Tank Level Control Simulation

Technologies

• LabVIEW Control Design and Simulation Module
  • http://www.ni.com/labview/cd-sim
• PID Theory
  • http://www.ni.com/white-paper/3782/en

Customer Benefits
Provided the customer with the ability to attain starting PID Gain values to implement in their application.

Solution
First, we created a mathematical model of simultaneously filling and draining a storage tank of cross section A. Then using the LabVIEW Control Design and Simulation Module, Rothenberg Industries created a simulation of the application to design the PID algorithm. After successfully simulation resists, we implemented the PID controller into the customer’s application.

We assume the following:

1. The liquid density is the same in the inlet, in the outlet, and in the tank.
2. The tank has straight, vertical walls.
3. Constant Control Volume
4. Pump Pressure constant throughout run
5. No Pump time delay
6. Steady Flow
7. For 30 psi to 200 psi, Density ≈ c, Spec Gravity ≈ 1

Figure 1: Constant volumetric flow in, PID controlled flow out

Flow through a valve is proportional to the square root of the pressure drop over the valve as follows.

For our control volume and surface, the conservation of linear momentum shows

Pressure and thus momentum are constant in the CV which yields

Where,

Therefore, the above equation can be rewritten as follows

For 30 psi-200 psi, ρ ≈ c, γ ≈ 1 yields

Bernoulli’s principle states for incompressible flows that

Where 1234 = 0, ℎ26783 = 0

Rearranging this in terms of gauge pressure gives the following

Substituting the gauge pressure into the above equation gives us,

Assuming ρ ≈ c, our mathematical model is as follows:

This is a nonlinear differential equation for h(t) and the block diagram for the model can be as shown in the Figure 2 below. This block diagram will then be programmed into the simulator VI. h(t) is numerically approximated (by the simulator) by integrating dh(t)/dt with respect to time, from time 0 to time t, with initial value h(0).

Figure 2: Open loop block diagram where the Control Valve is 100% open

Closed Loop Controller
We will now design the PID liquid level height controller. The controller will receive the liquid level height in real-time and control the variable output valve to meet the set reference height. A lookup table is used to determine the nonlinear Cv value from the OEM specifications.

Figure 3: Close-up diagram of the PID controller

Implementing the closed loop PID controller in the block diagram replaces the 100% open signal as shown below in Figure 4.

Figure 4: Closed loop block diagram where the Control Valve is PID Controlled

Using the LabVIEW Control Design and Simulation Module, we program our mathematical model into the simulation loop as shown below in Figure 5. The simulator will calculate and chart the liquid level h throughout the simulation. We use the PID Advanced vi for this controller.

Figure 5: Closed Loop Simulation

Configuring the simulation parameters as shown below in Figure 6 provides us with an approximate real time execution.

Figure 6: Simulation Settings

As you can see from the simulation above the water height starts at 80% with a setpoint of 25%. The output valve fully opens draining the water to match the setpoint height of 25%. At 5 seconds the setpoint increases to 50%, the output valve closes. The water height increases linearly as there is a constant volumetric input flow. As the water height reaches the setpoint the valve closes just enough to maintain the constant setpoint height.

 PID Autotuning & Gain Scheduling
To improve the performance of our controller lets include PID autotuning to tune our PID gain values. We will utilize the PID Relay technique which connects a relay and an extra feedback signal with the setpoint and keeps the PID controller in the loop with the relay. Uses the setpoint relay experiment to determine the information needed to tune the controller. Simulates an active step change (superimposed square wave) from Hset. Based on that response it determines the gains via the Ziegler and Nichols’ method.

Figure 8: Setpoint Relay Experiment

To include this into our LabVIEW program we will add the PID Online Autotuning vi and the PID Gain Schedule vi as show below in Figure 9. The gain schedule is a function of pump pressure. We created a gain schedule constant as well as several custom vis to Insert, Replace and Delete the tuned PID gains output by the autotuning vi into the PID gain schedule array.

Figure 9: Closed Loop Simulation with Autotuning and Gain Scheduling

So lets try out the latest simulation. Before you begin autotuning, you must establish a stable system, even if you cannot tune the system on your own. If the system model is too complex or nonlinear the autotune may not work. The algorithm assumes that the controlled system is first-order with a dead-time. 

Example closed loop simulation where V = 1000 m3, A = 20 m2, h0 = 80%, Cv = 0 - 0.75 m3Pa-1/2s-1, Q_in = 60 m3/s , P_T = 15 PSI

Where Hset = 25%  Initial Gains: Kc  10, Ti = 0.1

Once the system reaches steady state the Autotuning can be initaited. Tuned PI Gains are calculated as:

Kc = 18.39
Ti = 0.14
Td = 0.00

Figure 10: Autotuning Simulation

PI Autotuning Analysis:
Kc 10, Ti = 1

Figure 11: Controller Baseline

Kc = 18.39, Ti = 0.14

Figure 12: Controller with gains determined from Autotuning

Unfortunalty as you can see the Tuned gains do not yield much improvement. The documentation states, If the system model is too complex or nonlinear the autotune may not work.

Front Panel
Our final test front panel is shown below in Figure 13 below.

Figure 13: Front panel of the simulation vi